Method of Transmitting Data Using Space Time Block Codes

Abstract

To transmit data, a set of at least 2m n×n matrices that represent an extension of a fixed-point-free group is provided. To each of 2m of the matrices is allocated one of the binary numbers from 0 to 2m−1. The data are mapped into the matrices according to the allocation. The mapped matrices are transmitted, preferably using n antennas, one antenna per row of each matrix.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

The invention is herein described, by way of example only, with reference to the accompanying drawings, wherein:

FIGS. 1 and 2 are high-level block diagrams of transmitters of the present invention;

FIG. 3 is a high-level block diagram of a MIMO system;

FIG. 4 is a flow chart of simulation of the present invention;

FIG. 5 is the constellation of the quaternion group Q₂;

FIG. 6 is the constellation of the super-quaternion extension Q₂∪L₂∪L₄;

FIG. 7 is an example of bit allocation for an 8PSK gray code;

FIGS. 8-20 are plots of bit error rate vs. signal-to-noise ratio for various simulations of transmission using prior art space time block codes and space time block codes of the present invention, as discussed in the Theory Section;

FIG. 21 is the constellation of the super-quaternion extension of Table 1;

FIG. 22 is the constellation of the G_m,r coset extension of Table 2

FIG. 23 is a plot of bit error rate vs. signal-to-noise ratio for simulations of transmission using a prior art G_m,r space time block code and the G_m,r coset extension space time block code of Table 2.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

The present invention is of a method of transmitting digital data using space time is block codes. Specifically, the present invention can be used to transmit digital data with a lower bit error rate (BER) at a given signal-to-noise ratio (SNR) than according to the prior art.

The principles and operation of digital data transmission according to the present invention may be better understood with reference to the drawings and the accompanying description.

The theory of the present invention is described in the accompanying Theory Section. Two specific extensions of fixed-point-free groups are discussed in the Theory Section: super-quaternion extensions of quaternion groups and coset extensions of G_m,r groups. Table 1 lists an exemplary set of 2×2 matrix representations of the super-quaternion extension S_Q2,7,2, along with the associated bit allocations. Table 2 lists an exemplary set of 3×3 matrix representations of a coset extension of G_63,37, along with the associated bit allocations. The constellations of these two extensions are shown in FIGS. 21 and 22, respectively.

Referring now to the drawings, FIG. 1, which is based on the Figures of U.S. Pat. No. 6,185,258, is a high-level block diagram of a transmitter 10 of the present invention. Transmitter 10 includes a coder 12, a modulation and transmission section 14 and two antennas 16 and 18. Coder 12 receives an incoming stream of data bits to transmit and maps the bits, eight at a time, into the matrices of Table 1. For each matrix, modulation and transmission section 14 modulates a carrier wave according to the elements of the matrix and transmits the modulated carrier wave via antennas 16 and 18. First, modulation and transmission section 14 modulates the carrier wave according to the 1,1 matrix element and transmits the thus-modulated carrier wave via antenna 16, while also modulating the carrier wave according to the 2,1 matrix element and transmitting the thus-modulated carrier wave via antenna 18. Then, modulation and transmission section 14 modulates the carrier wave according to the 1,2 matrix element and transmits the thus-modulated carrier wave via antenna 16, while also modulating the carrier wave according to the 2,2 matrix element and transmitting the thus-modulated carrier wave via antenna 18.

For example, to transmit the bit pattern “00100011” (binary 67), the matrix

$\hspace{1em}\left(\begin{array}{cc}-0.8944i& -0.4472\\ 0.4472& 0.8944i\end{array}\right)$

is transmitted. First, modulation and transmission section 14 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.8944 and shifting the phase of the carrier wave by −90°, and transmits the thus-modulated carrier wave via antenna 16. Simultaneously, modulation and transmission section 14 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.4472 without shifting the phase of the carrier wave, and transmits the thus-modulated carrier wave via antenna 18. Then, modulation and transmission section 14 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.4472 and shifting the phase of the carrier wave by 180°, and transmits the thus-modulated carrier wave via antenna 16. Simultaneously, modulation and transmission section 14 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.8944 and shifting the phase of the carrier wave by 90°, and transmits the thus-modulated carrier wave via antenna 18.

At the receiver, the received signal is demodulated and conventional maximum likelihood estimation is used to estimate the transmitted matrix. Then the bit pattern associated with the closest matrix, in a look-up table identical to Table 1, to the estimated matrix is interpreted as the transmitted bit pattern.

The receiver may receive the signal via a known channel or via an unknown channel. In a known channel, the receiver knows the channel response from every transmitter antenna to every receiver antenna. For example, in some channels, before the transmission of a block of data, a predefined preamble is transmitted. The receiver learns the characteristics of the channel from the preamble. If the channel characteristics change much more slowly than the time it takes to transmit a block of data, the channel is considered “known”. Otherwise, the channel is considered “unknown”. Both cases are treated in the Theory Section.

FIG. 10 shows the bit error rate (BER) obtained in several simulations of transmitting using super-quaternion extensions of quaternion groups, vs. signal-to-noise ratio (SNR), compared with simulations of transmitting using two prior art space time block codes. The results using the super-quaternion extension of Table 1 are labeled “Rate 4 Super Quaternion” in FIG. 10.

An alternative transmitter 30 of the present invention is illustrated in FIG. 2. Transmitter 30 includes a coder 32, a modulation and transmission section 34 and three antennas 36, 38 and 40. Coder 32 receives an incoming stream of data bits to transmit and maps the bits, nine at a time, into the matrices of Table 2. For each matrix, modulation and transmission section 34 modulates a carrier wave according to the elements of the matrix and transmits the modulated carrier wave via antennas 36, 38 and 40. First, modulation and transmission section 34 modulates the carrier wave according to the 1,1 matrix element and transmits the thus-modulated carrier wave via antenna 36, while also modulating the carrier wave according to the 2,1 matrix element and transmitting the thus-modulated carrier wave via antenna 38, and while also modulating the carrier wave according to the 3,1 matrix element and transmitting the thus-modulated carrier wave via antenna 40. Then, modulation and transmission section 34 modulates the carrier wave according to the 1,2 matrix element and transmits the thus-modulated carrier wave via antenna 36, while also modulating the carrier wave according to the 2,2 matrix element and transmitting the thus-modulated carrier wave via antenna 38, and while also modulating the carrier wave according to the 3,2 matrix element and transmitting the thus-modulated carrier wave via antenna 40. Finally, modulation and transmission section modulates the carrier wave according to the 1,3 matrix element and transmits the thus-modulated carrier wave via antenna 36, while also modulating the carrier wave according to the 2,3 matrix element and transmitting the thus-modulated carrier wave via antenna 38, and while also modulating the carrier wave according to the 3,3 matrix element and transmitting the thus-modulated carrier wave via antenna 40.

For example, to transmit the bit pattern “100111000” (binary 312), the matrix

$\hspace{1em}\left(\begin{array}{ccc}-0.4460-0.1866i& -0.5437+0.4109i& 0.3838-0.3929i\\ -0.05577+0.3449i& 0.0970+0.1204i& 0.2427+0.6980i\\ 0.1845-0.5497i& 0.5460+0.4621i& 0.3655+0.1364i\end{array}\right)$

is transmitted. First, modulation and transmission section 34 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.4835 (the absolute value of the 1,1 matrix element) and shifting the phase of the carrier wave by −156.0° (the phase of the 1,1 matrix element), and transmits the thus-modulated carrier wave via antenna 36. Simultaneously, modulation and transmission section 34 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.6557 (the absolute value of the 2,1 matrix element) and shifting the phase of the carrier wave by 148.3° (the phase of the 2,1 matrix element), and transmits the thus-modulated carrier wave via antenna 38. Also simultaneously, modulation and transmission section 34 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.5798 (the absolute value of the 3,1 matrix element) and shifting the phase of the carrier wave by −71.4° (the phase of the 3,1 matrix element), and transmits the thus-modulated carrier wave via antenna 40. Then, modulation and transmission section 34 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.6815 (the absolute value of the 1,2 matrix element) and shifting the phase of the carrier wave by 142.9° (the phase of the 1,2 matrix element), and transmits the thus-modulated carrier wave via antenna 36. Simultaneously, modulation and transmission section 34 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.1546 (the absolute value of the 2,2 matrix element) and shifting the phase of the carrier wave by 51.1° (the phase of the 2,2 matrix element), and transmits the thus-modulated carrier wave via antenna 38. Also simultaneously, modulation and transmission section 34 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.7153 (the absolute value of the 3,2 matrix element) and shifting the phase of the carrier wave by 40.20 (the phase of the 3,2 matrix element), and transmits the thus-modulated carrier wave via antenna 40. Finally, modulation and transmission section 34 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.5493 (the absolute value of the 1,3 matrix element) and shifting the phase of the carrier wave by ˜45.7 (the phase of the 1,3 matrix element), and transmits the thus-modulated carrier wave via antenna 36. Simultaneously, modulation and transmission section 34 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.7390 (the absolute value of the 2,3 matrix element) and shifting the phase of the carrier wave by 70.8° (the phase of the 2,3 matrix element), and transmits the thus-modulated carrier wave via antenna 38. Also simultaneously, modulation and transmission section 34 modulates the carrier wave by multiplying the amplitude of the carrier wave by 0.3501 (the absolute value of the 3,3 matrix element) and shifting the phase of the carrier wave by 20.5° (the phase of the 3,3 matrix element), and transmits the thus-modulated carrier wave via antenna 40.

At the receiver, the received signal is demodulated and conventional maximum likelihood estimation is used to estimate the transmitted matrix. Then the bit pattern associated with the closest matrix, in a look-up table identical to Table 2, to the estimated matrix is interpreted as the transmitted bit pattern.

FIG. 23 shows the BER obtained in a simulation of transmitting using the G_m,r coset extension of Table 2 vs. SNR, compared with a simulation of transmitting using a prior art G_m,r group.

While the invention has been described with respect to a limited number of embodiments, it will be appreciated that many variations, modifications and other applications of the invention may be made.

Theory Section

Abstract

Space Time Block Codes (STBC) are designed for Multiple Input—Multiple Output (MIMO) channels. The recently developed advanced codes (Turbo codes and LDPC) have been able to achieve rates at almost the channels' capacity in Single Input—Single Output (SISO) channels. Given a fixed transmit power and bandwidth it is necessary to take advantage of space diversity if one wishes to exceed the Shannon limit for data rate. Furthermore, in order to avoid errors, SISO fading channels require long coding blocks and interleavers that result in high delays. STBC codes take advantage of space-time diversity to reduce errors. Joined STBC with error correction codes can achieve high rates with low Bit Error Rate (BER) at low delays.

Early STBC, that where developed by Alamouti [2] for known channels and by Tarokh [3] for unknown channels, have been proven to increase the performance of channels characterized by Rayleigh fading. It has been shown in [4,6] that the diversity of STBC codes is a good criterion for its performance. Codes that are based on groups or division algebras have by definition non-zero diversity and therefore they are suitable for STBC in order to achieve high rates at low Symbol to Noise Ratio (SNR). This work presents a new high diversity group based on STBC with improved performance both in known and unknown channels.

1. INTRODUCTION

Current communications data rates increase rapidly in order to support endpoint application requirements for ever growing bandwidth and communication speed. While 1st generation cellular equipment supported up to 9.6 kbps data rate, 2nd generation cellular supported up to 57.6 kbps rates, and today, 3rd generation standards support up to 384 kbps rates (local area coverage of up to 2 Mbps). The WLAN standard 802.11a supports up to 54 Mbps and 802.11b supports up to 11 Mbps, while the maximal data rate of the 802.11 standard is 3 Mbps.

Although the rates of the communication standards are increasing, the transmit power remains low due to radiation limitations and collocation requirements. Naturally, higher data rates with fixed transmit power result in range degradation. Since Shannon's limit has almost been achieved by recently developed error correction codes, it is necessary to exceed SISO channels performance by taking advantage of space diversity. FIG. 3 illustrates a typical MIMO system.

2. PREVIEW ON STBC

In this section we give a short preview on STBC and the motivation for using receive and transmit diversity.

2.1. STBC Constellation

The STBC codes, that we presented here are based on constellations of the form α·e^iβ=a+i·b, where α is the amplitude of the modulation and β is the phase of the modulation. When we refer to a code we mean a set of matrices, where each matrix is a codeword.

The matrices represent n² constellations each (for n-transmit diversity) where the column represents the transmission antenna and the row represents the time. If the matrix {c_ij}_i,j=1ⁿ is transmitted at times T, T+1, . . . , T+n−1 then the symbol c_ij will be transmitted at time i from antenna j. The following example illustrates the transmission of a code word c_ij for 2-transmit diversity (n=2) at times T, T+1:

Time	Antenna 1	Antenna 2
T	C11	C12
T + 1	C21	C22

Transmit diversity refers to multiple transmitter antennas, and receive diversity refers to multiple receiver antennas. The STBC codes are usually designed according to the Transmit diversity, while receive diversity has merely effect on the performance. The receiver estimates the transmitted codeword according to n (n is the number of transmitter antennas) time samples from the receiver antennas.

We consider two channel types:

• Known Channel In a known channel the receiver knows the channel response from every transmitter antenna to every receiver antenna. For example, in some channels before the transmission of every data block, a predefined preamble is transmitted and the receiver learns the channel from the preamble. If the channel changes much slower than the data block length then the channel is considered to be known.
• Unknown Channel In this case the receiver has no knowledge of the channel, and the data has to be decoded differentially as shown in section 9.

2.2 Channel Model

Multipath channels in urban and suburban areas are very difficult to estimate. Beside the specific dependency on place, speed, and surrounding moving objects (cars, people) the channels characteristics are specific to topography, building style and density, frequency, antenna height, type of terrain and other parameters. Small change in one of these parameters may affect completely the channel response. Many channel response simplification models have been introduced to estimate the channel performance. Some of the most commonly used models are given below.

• • Free Space The only accurate model. The path loss L (power loss due to channel response) in this case depends entirely on the distance R (and on the atmosphere at frequencies above 20 Ghz):

$L\propto \frac{1}{R^2}.$

• • Flat Land: Mathematical estimation of a path loss for a line of sight channels over flat land with low antennas (the height of antennas is much smaller than the distance between them). The loss is given by

$L\propto \frac{1}{R^4}.$

• • Egli: This propagation model takes into account experimental dependency of the path loss on the frequency f:

$L\propto \frac{1}{R^4f^2}$

• • TIREM, ITU-R: Experimental models that give path loss estimation based on field experiments.
  • Rice: A random model for a line of sight channel that involves multipath.
  • Rayleigh: A random model for multipath channels without line of sight. The channel's response in this model (for a SISO channel) is

$\begin{array}{cc}Y=\alpha ·\frac{1}{\sqrt{2}}\left({\upsilon }_1+i{\upsilon }_2\right)X& \left(1\right)\end{array}$

• •  where X is the transmitted symbol, Y the received symbol, α is a constant that equals the average absolute path loss, and ν₁,ν₂ are i.i.d. normal Gaussian variables.

The model used in our simulations is the Rayleigh model [12], which is adequate for urban communications with no line of sight. The channel's response of a MIMO channel (FIG. 3) is represented by the matrix

$R=\left(\begin{array}{cccc}{r}_{11}& r_{12}& \cdots & r_{1m}\\ r_{21}& r_{22}& \cdots & r_{2m}\\ ⋮& ⋮& ⋰& ⋮\\ r_{n1}& r_{n2}& \cdots & r_{nm}\end{array}\right)$

where r_ij is the path loss from transmitter antenna i to receiver antenna j, i=1, . . . , n, j=1, . . . , m, n is the transmit diversity, and m is the receive diversity. r_ij are assumed to be independent (this is usually the case when all the antennas are placed more than half wavelength from each other) and the distribution of r_ij is given by EQ. 1.

The channel's noise at the receiver antennas is assumed to be i.i.d normal Gaussian noise, with random, uniformly distributed phase. Since the total transmission power over all the transmitter's antennas is normalized to 1, the average SNR at the receiver antennas is a from EQ. 1, given in dB.

The new proposed algorithms were simulated to estimate their performance. The simulations were performed over a large number of independent tests to assure accurate results. All symbols/bits at the bit/symbol generator were chosen randomly and independently and the channel noise was generated as i.i.d. normal Gaussian noise for all time samples and receiver antennas. At the known channel simulations the channel response was generated independently for each codeword and was also delivered to the receiver as side information. At the unknown channel simulations the channel response was generated for a block of codewords, while the first codeword in every block was transmitted as reference and was not estimated in order to eliminate simulation dependency (for large enough number of tests) on block length. The bit allocation was decided prior to the simulation according to numerical optimization simulations. In cases bit allocation was not possible the symbols were generated directly and the BER was estimated according to chapter 7. The flow of the simulation scheme is given in FIG. 4.

• • Random Bit Generator: Simulates the information bits of an information source for codes with integral rate. The bits are equally distributed and independent.
  • Bit Allocation: This procedure allocates the bits into channel symbols. The bits are allocated optimally to achieve minimal error probability. The algorithm allocates closer symbols to codewords with smaller hamming distance (section 7).
  • Symbol Generator: In the case of codes with a non-integral rate the symbols are generated directly with no bit allocation. All symbols are i.i.d.
  • Space Time Realization: This procedure represents the codewords as matrices, as described in section 2.1.
  • Channel Response Generator: Simulates Rayleigh fading channel. Every transmitted symbol is multiplied by the channel response (elaborated at section 2.2). The variance of the channel's response is calculated according to the SNR (EQ. 1).
  • Channel Noise Generator: Gaussian white noise is added to the receiver antennas. The variance of the noise is set to 1.
  • Receiver realization: This procedure calculates the received symbols for every receiver antenna according to the transmitted codewords, the channel response and the channel noise.
  • Maximum Likelihood Estimator: For every Block of received symbols (the size of the block is the number of transmitter antennas) we estimate the maximum likelihood transmitted codeword. When known channel is simulated all the channel response parameters are given to the ‘Maximum Likelihood Estimator’ as side information. When unknown channel is simulated the estimation is made differentially according to the previous received symbols (section 9).
  • Bit Translation: The symbols are translated back into bits (for codes with whole rates).
  • Comparator: The BER is calculated by comparing the transmitted symbols/bits to the received symbols/bits and averaging over a large number of samples. The calculations of BER are elaborated in section 7.

2.3. Diversity and Error Probability

In this section we will assume that the receiver has perfect knowledge of the channel (i.e. in noiseless channel there would have been a perfect reconstruction). We assume that the number of transmitters is n and the number of receivers is m. Given that the codeword

{right arrow over (c)}=c₁¹c₁² . . . c₁ⁿc₂¹c₂² . . . c₂ⁿc₃¹ . . . c_n¹ . . . c_nⁿ

was transmitted, then the received signal,

{right arrow over (d)}=d₁¹d₁² . . . d₁^md₂¹d₂² . . . d₂^md₃¹ . . . d_n¹ . . . d_n^m

is given by

$d_t^j=\sum _{i=1}^n\left({\alpha }_{i,j}c_t^i\sqrt{E_s}\right)+v_t^j,1\le j\le m$

where j is the receiver antenna and t is the time sample, α_i,j is the channel response from transmitter i to receiver j, c_tⁱ is the transmitted signal from transmitter i at time t, E_s is the normalization factor for the transmission power in order to keep the average constellation equal to 1, v_t^j is a Gaussian white noise at the receiver antenna.

Consider that the following signal

{right arrow over (e)}=e₁¹e₁² . . . e₁ⁿe₂¹e₂² . . . e₂ⁿe₃¹ . . . e_k¹ . . . e_kⁿ

is the receiver's decision upon maximum-likelihood criteria.

When white Gaussian noise is present, the probability of deciding in favor of e, when c was transmitted is ([6,4]):

$P\left(\overrightarrow{c}\overrightarrow{e}|{\alpha }_{i,j}\right)=\frac{1}{\sqrt{\pi N_0}}{\int }_{d\left(c,e\right)\le d\left(c,e^{\prime }\right)}^{\infty }{}^{-\frac{d^2\left(c,e^{\prime }\right)}{N_0}}e^{\prime }\le {}^{-d^2\left(c,e\right)E_s/4N_0},1\le i\le n,1\le j\le m$

where N₀/2 is the noise variance per receiver's element and

$d^2\left(c,e\right)=\sum _{j=1}^m\sum _{t=1}^k{\sum _{i=1}^n{\alpha }_{i,j}\left(c_j^i-e_t^i\right)}^2.$

Define Ω_j(α_1,j, . . . , α_n,j). We get after some manipulations that

$d^2\left(\overrightarrow{c},\overrightarrow{e}\right)=\sum _{j=1}^m{\Omega }_jA\left(\overrightarrow{c},\overrightarrow{e}\right){\Omega }_j^{*}$ $\mathrm{where}$ $A_{\mathrm{pq}}=\left(c_1^p-e_1^p,c_2^p-e_2^p,\dots ,c_k^p-e_k^p\right){\left(c_1^q-e_1^q,c_2^q-e_2^q,\dots ,c_k^q-e_k^q\right)}^{*}$

and * is the conjugate-transpose (conjugate for scalars). Thus, the error probability can be estimated to be

$\begin{array}{cc}P\left(\stackrel{\_}{c}->\stackrel{\_}{e}{\alpha }_{i,j}\right)\le \prod _{j=1}^m{}^{-{\Omega }_jA\left(\stackrel{\_}{c},\stackrel{\_}{e}\right){\Omega }_j^{*}E_s/4N_0},1\le i\le n,1\le j\le m.& \left(2\right)\end{array}$

Since A({right arrow over (c)}, {right arrow over (e)}) is a Hermitian matrix there exists a unitary matrix V, which consists of the eigenvectors of A, such that VA(c,e)V*=D is diagonal. D_i,i=Λ_i are the eigenvalues. Clearly, the matrix

$B\left(\stackrel{\_}{c},\stackrel{\_}{e}\right)=\left(\begin{array}{cccc}{c}_1^1-e_1^1& c_2^1-e_2^1& \dots & c_k^1-e_k^1\\ ⋮& ⋮& ⋰& ⋮\\ c_1^n-e_1^n& c_2^n-e_2^n& \dots & c_k^n-e_k^2\end{array}\right).$

is a square root of A(c,e), i.e A({right arrow over (c)},{right arrow over (e)})=B({right arrow over (c)},{right arrow over (e)})*B({right arrow over (c)},{right arrow over (e)}).

Let

$\begin{array}{cc}\stackrel{\_}{B_j}\stackrel{\Delta }{=}\left(B_1,j,\dots ,{\beta }_{n,j}\right)={\Omega }_jV^{*}& \left(3\right)\\ \mathrm{then}& \\ {\Omega }_jA\left(\stackrel{\_}{c},\stackrel{\_}{e}\right){\Omega }_j^{*}=\sum _{i=1}^n{\lambda }_i{{\beta }_{i,j}}^2.& \end{array}$

Now we can estimate the probability of EQ. 2 by:

$\begin{array}{cc}P\left(\stackrel{\_}{c}->\stackrel{\_}{e}\left\{a_{i,j}\right\}\right)\le \prod _{j=1}^m{}^{\frac{-E_s}{4N_0}\sum _{i=1}^n{\lambda }_i{{\beta }_{i,j}}^2},1\le i\le n,1\le j\le m.& \left(4\right)\end{array}$

Assume we have an independent channel response among different paths (i.e. α_i,j are independent). Since V is unitary, it is easy to show using definition 3 that β_i,j are i.i.d. After averaging over β_i,j, we get for Rayleigh fading the following expression:

$P\left(\stackrel{\_}{c}->\stackrel{\_}{e}{\alpha }_{i,j}\right)\le {\left(\frac{1}{\prod _{i=1}^n\left(1+{\lambda }_iE_s/4N_0\right)}\right)}^m,1\le i\le n,1\le j\le m.$

We can assume for high SNR that 1+λ_iE_s/4N₀ λ_iE_s/4N₀, then

$\begin{array}{cc}P\left(\stackrel{\_}{c}->\stackrel{\_}{e}{\alpha }_{i,j}\right)\le {\left(\prod _{i=1}^r{\lambda }_i\right)}^{-m}{\left(E_s/4N_0\right)}^{-rm},1\le i\le n,1\le j\le m& \left(5\right)\end{array}$

where r is the number of non-zero eigenvalues of A.

By assuming that A is regular (r=n), we get

$\prod _{i=1}^r{\lambda }_i=\det \left(A\right),$

but |A|=|B|² and therefore

P({right arrow over (c)}→{right arrow over (e)}|α_i,j)≦|B|^−2m(E_x/4N₀)^−rm, 1≦i≦n, 1≦j≦m.

Since B is the difference between two distinct signals, c and e, we define the diversity to be

$\zeta \stackrel{\Delta }{=}\frac{1}{2}\min {\det \left(C_l-C_{l^{\prime }}\right)}^{\frac{1}{m}}$

for all distinct code words C_l and C_l′. It is obvious that codes that assure maximal diversity will achieve smaller error probabilities.

2.4. Receive and Transmit Diversity

The STBC codes are designed to take advantage of the transmit diversity. However, it is a well known fact that the capacity and the performance of a MIMO channel depends on the receive diversity. The standard formula for Shannon capacity is given by:

C=log₂(1+ρ|H|²)  (6)

where ρ is the transmission power (normalized to the channel noise) and H is the channel transfer power characteristic. Assume that the transmitted signals at the transmitter's antennas are independent, have equal power and mutually Gaussian distributed. The capacity of a MIMO channel [12] is:

C=log₂[det(I_N+(ρ/N)HH*)] bps/Hz.  (7)

For example, for N independent deterministic channels, i.e. H=I_N, we get

$\begin{array}{cc}\underset{N->\infty }{\lim }C_N=\underset{N->\infty }{\lim }N·{\log }_2\left(1+\left(\rho /N\right)\right)=\rho /\ln \left(2\right).& \left(8\right)\end{array}$

The capacity in this case is linear with ρ, while the dependency is logarithmic in Eq. 6. Next we consider the case of M receiver antennas and a single transmitter antenna. The capacity is:

$\begin{array}{cc}C={\log }_2\left(1+\rho \sum _{i=0}^M{H_i}^2\right).& \left(9\right)\end{array}$

For a Rayleigh fading channel the capacity of this channel is:

$\begin{array}{cc}C={\log }_2\left(1+\rho ·{\chi }_{2M}^2\right).& \left(10\right)\\ \mathrm{where}& \\ {\chi }_{2M}^2\propto \sum _{i=0}^{2M}X_i^2\mathrm{and}X_i\mathrm{are}i.i.d,X_i=\mathrm{NORMAL}\left(0,1/\sqrt{2}\right).& \end{array}$

The capacity of a MIMO channel with N transmitter antennas and a single receiver antenna in a fading channel is

C=log₂(1+ρχ₂²).  (11)

The inferior capacity of transmit diversity, compared to receive diversity, is due to the constraint on the overall transmit power.

The capacity of a SISO fading channel is

C=log₂(1+ρχ₂²).  (12)

Due to the convexity of the log function the MIMO channel with transmit diversity has in average a larger capacity than the SISO channel.

From EQ. 5 we can see that the error probability depends exponentially on the number of receiver antennas, for a fixed transmit diversity. FIG. 20 shows the improved performance when two receiver antennas are used.

3. PRIOR RELATED WORK

In this section we present a short review of known STBC constellations that are based on algebraic structures. The new STBC codes, developed in this invention, resemble some of the codes presented in this section.

3.1. Usage of Group Structures for Space Time Coding

The usage of groups to design unitary space-time constellations has high potential to provide constellations with good diversity product. We will review some of these groups which were analyzed in [8].

Cyclic Groups:

Cyclic groups are used for differential modulation in [7], and also referred to as “diagonal designs”. The elements of these groups are diagonal Lth roots of unity.

V_l=V_s^l and V_s=diag[e^i2πu1/L, . . . , e^i2πum/L].

The u_m are chosen to maximize the diversity product ζ_V, which is given by

$\begin{array}{c}{\zeta }_V=\frac{1}{2}\underset{0l<l^{\prime }<L}{\min }{\det \left(V_l-V_{l^{\prime }}\right)}^{\frac{1}{M}}\\ =\frac{1}{2}\underset{l=1,\dots ,L-1}{\min }{\det \left(I-V_l\right)}^{\frac{1}{M}}\\ =\frac{1}{2}\underset{l=1,\dots ,L-1}{\min }{\det \left(\mathrm{diag}\left[1-{}^{2\pi u_i/L},\dots ,1-{}^{2\pi u_m/L}\right]\right)}^{\frac{1}{M}}\\ =\frac{1}{2}\underset{l=1,\dots ,L-1}{\min }{\prod _{i=1}^M1-{}^{2\pi u_i/L}}^{\frac{1}{M}}\\ =\frac{1}{2}\underset{l=1,\dots ,L-1}{\min }{\prod _{i=1}^M{}^{\pi u_i/L}\left({}^{-\pi u_i/L}-{}^{\pi u_i/L}\right)}^{\frac{1}{M}}\\ =\underset{l=1,\dots ,L-1}{\min }{\prod _{i=1}^M\frac{{}^{\pi u_i/L}-{}^{-\pi u_i/L}}{2}}^{\frac{1}{M}}\\ =\underset{l=1,\dots ,L-1}{\min }{\prod _{i=1}^M\sin \frac{\pi u_il}{L}}^{\frac{1}{M}}.\end{array}$

In this constellation, the transmitter's antennas are activated one at a time, sequentially, and in the same order. Notice that these groups are Abelian (commutative), and that they have full diversity, i.e. positive diversity. Diagonal designs perform well in low rates, but other works such as in [8] tried to find designs with matrices that are not “sparse” and will achieve good diversity at high rates.

Quaternion Groups:

The quaternion groups are also called “dicyclic groups” in [9], and have the form

Q _p=<α,β|+^2p=1,β²=α^2p−1,βα=α⁻¹β>, p≧1

where <·> refers to the groups generated from the elements that are in the brackets. The group order is L=2^p+1. The matrix representation for the 2×2 quaternion group appears in [9], so the rate of the constellations will be R=(p+1)/2 and the group is generated from two unitary matrices:

$〈\left(\begin{array}{cc}{}^{2\pi i/2^p}& 0\\ 0& {}^{-2\pi i/2^p}\end{array}\right),\left(\begin{array}{cc}0& 1\\ -1& 0\end{array}\right)〉.$

These groups will be used in our work for the development of new constellations, which are not necessarily groups, but have high diversity, and exceedingly good performance at low SNR.

Fixed Point Free Groups:

A necessary condition for a finite group of matrices to have full diversity (ζ>0) is that it be fixed point free. These types of groups were explored by Burnside, Zassenhaus, Amitsur and others. Cyclic groups and quaternion groups are specific examples of such groups. A recent comprehensive survey is [8]. There are six types of fix-point free groups: G_m,r, D_m,r,l, E_m,r, F_m,r,l, J_m,r and K_m,r,l. G_m,r, for example, has the form of

G_m,r=<a,b|a^m=1,bⁿ=a^t,ba=a^rb>  (13)

where n is the order of r modulo m (i.e. n is the smallest positive integer such that rⁿ≡1 mod m), t=m/gcd(r−1,m), and gcd(n,t)=1. The matrix representation of G_m,r is A^sB^k, s=0, . . . , m−1, k=0, . . . , n−1 where

$\begin{array}{cc}A=\left(\begin{array}{ccccc}\xi & 0& 0& \dots & 0\\ 0& {\xi }^r& 0& \dots & 0\\ 0& 0& {\xi }^{r^2}& \dots & 0\\ ⋮& ⋮& ⋮& ⋰& ⋮\\ 0& 0& 0& \dots & {\xi }^{r^{n-1}}\end{array}\right),B=\left(\begin{array}{ccccc}0& 1& 0& \dots & 0\\ 0& 0& 1& \dots & 0\\ ⋮& ⋮& ⋮& ⋰& ⋮\\ 0& 0& 0& \dots & 1\\ {\xi }^t& 0& 0& \dots & 0\end{array}\right)& \left(14\right)\\ \mathrm{and}\xi ={}^{2\pi i/m}.& \end{array}$

It can be proved that a finite group is fixed-point-free if and only if it is isomorphic to one of the above six groups [8]. Many of the fixed-point-free group representations have good diversity product and therefore provide high rate constellations.

3.2. Extensions for Group-Related Structures

Although we saw that group constellations have better chance to have full diversity, sometimes the group constraint does not really help, and we may want to look at matrix sets that are not necessarily groups. Moreover, there is a strong motivation to develop group-related structures, that still have small number of distinct pairwise distances.

Orthogonal Designs:

Orthogonal designs were introduced in the early stages of space-time codes development (see [5,2]). A complex orthogonal design of size N is an N×N unitary matrix whose rows are permutations of complex numbers ±x₁, ±x₂, . . . , ±x_n, their conjugates ±x₁*, ±₂*, . . . , ±x_n*, or multiples of these indeterminates by ±√{square root over (−1)}. The matrix representation of orthogonal designs for two transmitter antennas is:

$\begin{array}{cc}OD\left(x,y\right)=\frac{1}{\sqrt{2}}\left(\begin{array}{cc}x& -y^{*}\\ y& x^{*}\end{array}\right)& \left(15\right)\end{array}$

where x and y are subject to a power constraint. In order to have unitary matrices we can use the constraint that |x|²=|y|²=1. The unitary constellations are then obtained by letting x,y be the Qth roots of unity, so

V={OD(x,y)|x,yε{1,e^2πi/Q, . . . , e^{2πi(Q−1)/Q}}}.  (16)

These constellations do not generally form a group, but perform well as space-time block codes [5].

Nongroup Generalizations of G_m,r:

As shown in this section G_m,r (Eq. 6) has a matrix representation of dimension n, where n is a function of m and r. We can try and relax the constraint on the elements of the matrices that form G_m,r and look at the general case of the set S_m,s, consisting of the matrices A^lB^k where l=0, . . . , m−1, k=0, . . . , min(s,n),

$\begin{array}{cc}A=\left(\begin{array}{ccccc}{\alpha }^{u_1}& 0& 0& \dots & 0\\ 0& {\alpha }^{u_2}& 0& \dots & 0\\ 0& 0& {\alpha }^{u_3}& \dots & 0\\ ⋮& ⋮& ⋮& ⋰& ⋮\\ 0& 0& 0& \dots & {\alpha }^{u_n}\end{array}\right),B=\left(\begin{array}{ccccc}0& 1& 0& \dots & 0\\ 0& 0& 1& \dots & 0\\ ⋮& ⋮& ⋮& ⋰& ⋮\\ 0& 0& 0& \dots & 1\\ \beta & 0& 0& \dots & 0\end{array}\right)& \left(17\right)\end{array}$

where α,β are the primitive mth and sth roots of unity, and u₁, . . . , u_n are integers. Deeper analysis of these constellations can be found in [8].

Products of Group Representations:

Consider two fixed-point free groups S_A and S_B and their unitary M×M matrix representations {A₁, . . . , A_LA} and {B₁, . . . , B_LB}, respectively. Now we look at the set of pairwise product

S_A,B={A_jB_k|j=1, . . . , L_A, k=1, L_B}.  (18)

S_A,B has at most L_AL_B distinct elements, and the constellation rate is at most R=R_A+R_B. The diversity product is then:

$\begin{array}{c}{\zeta }_s=\frac{1}{2}\underset{\left(j,k\right)\ne \left(j^{\prime },k^{\prime }\right)}{\min }{\det \left(A_jB_k-A_{j^{\prime }}B_{k^{\prime }}\right)}^{\frac{1}{M}}\\ =\frac{1}{2}\underset{\left(j,k\right)\ne \left(j^{\prime },k^{\prime }\right)}{\min }{\det \left(A_{j^{\prime }}^{-1}\right)\det \left(A_jB_k-A_{j^{\prime }}B_{k^{\prime }}\right)\det \left(B_k^{-1}\right)}^{\frac{1}{M}}\\ =\frac{1}{2}\underset{\left(j,k\right)\ne \left(j^{\prime },k^{\prime }\right)}{\min }{\det \left(A_{j^{\prime }}^{-1}A_j-B_{k^{\prime }}B_k^{-1}\right)}^{\frac{1}{M}}\\ =\frac{1}{2}\underset{\left(l,l^{\prime }\right)\ne \left(0,0\right)}{\min }{\det \left(A_l-B_{l^{\prime }}\right)}^{\frac{1}{M}}\end{array}$

Notice that we use the fact that the determinant of the unitary matrix equals 1, and that S_A and S_B are groups. Finally, we see that even though S_A,B is not a group, it has at most L−1, rather then L(L−1)/2, distinct pairwise distances. Therefore, it has a good chance of having full diversity. Deeper analysis of group products can be found in [8].

3.3. Projective Groups and Their Performances

In order to achieve high performance space-time coding we design a code that is based on unitary matrices that provide us with good diversity product values. One of the structure that we explore is the projective groups. A projective group is a collection of elements V that satisfies the property that if A,BεV, then AB=αC where CεV and α is a scalar. In other words, the product of every two elements in V, is also an element of V, multiplied by a scalar. Projective groups have a less strict requirement to satisfy for a collection of signal matrices, but can still maintain the decoding complexity criterion. This is because for differential space time modulation, the matrix multiplication can be replaced by a group lookup table. For example, S_a1 S_a2 S_a3 S_a4=α₁α₂α₃S_a5.

We design a set of matrices that form a projective group by the following construction: let α,β be some non rational numbers, and let ζ=e^2πi/3. The elements of the projective group P_α,β,3 are given by AⁱB^j where j=0, 1, 2, i=0, 1, 2 and

$B=\left(\begin{array}{ccc}0& 1& 1\\ 0& 0& 1\\ \beta & 0& 0\end{array}\right),A=\left(\begin{array}{ccc}\sqrt[3]{\alpha }& 0& 0\\ 0& \zeta \sqrt[3]{\alpha }& 0\\ 0& 0& {\zeta }^2\sqrt[3]{\alpha }\end{array}\right).$

The relation between α,β and the diversity product ζseems to be chaotic. Some of the diversity products for different α and β are given in the following table. In this table. R is the transmission rate, L is the size of the group, M is the number of transmitting antennas and ζ is the diversity product.

	R	L	M	ζ	Group structures
	1.06	9	3	0.3971	Pα,β,3 with
					α = e2πi{square root over (5)}, β = e2πi{square root over (11)}
	1.06	9	3	0.4273	Pα,β,3 with
					α = e2πi{square root over (6)}, β = e2πi{square root over (32)}
	1	4	2	0.3741	Pα,β,2 with
					α = e2πi{square root over (5)}, β = e2πi{square root over (7)}
	1	4	2	0.7053	Pα,β,2 with
					α = e2πi{square root over (42)}, β = e2πi{square root over (90)}

4. EXTENSIONS OF THE QUATERNION GROUPS—NEW STBC

This section presents new STBC codes, that are based on extensions of the quaternion groups and the performance of STBC, both for known and unknown channels.

4.1. Super Quaternion Sets

When we examine the diversity product of the matrices in the constellations that form a group with respect to matrix multiplication, we realize that the performance of the quaternion group outperforms other structures. This fact led us to search extended sets of quaternion that have more elements while preserving a high diversity product. We may lose the group structure in this construction. However, if we think of the diversity product as the minimum distance between the elements of the constellation, we can design a structure where the elements are “well-spaced”, in the sense that the distance between every two elements is sufficiently big.

We begin with 2×2 matrix representation of the quaternion group Q₂. It has eight elements. The transmission rate is R=log₂ 8/2 and the diversity product is ζ_Q2=0.7071. The group is constructed from the set of matrices ±1, ±Q_i, ±Q_j, ±Q_k where:

$1=\left(\begin{array}{cc}1& 0\\ 0& 1\end{array}\right)Q_i=\left(\begin{array}{cc}0& 1\\ -1& 0\end{array}\right)Q_j=\left(\begin{array}{cc}i& 0\\ 0& -i\end{array}\right)Q_k=\left(\begin{array}{cc}0& -i\\ -i& 0\end{array}\right).$

We now define the super-quaternion of Q₂ with n layers (explained later), for matrices of degree 2, as a set of all linear combinations of the matrices Q_i that satisfy:

$\begin{array}{cc}{S}_{Q_2,n,2}=\left\{\frac{\begin{array}{c}{x}_1Q_i+x_2Q_j+\\ x_3Q_k+x_4Q_l\end{array}}{\sqrt{x_1^2+x_2^2+x_3^2+x_4^2}}\begin{array}{c}0<x_1^2+x_2^2+x_3^2+x_4^2\le n,\\ x_i\lambda \in Z\end{array}\right\}.& \left(19\right)\end{array}$

The matrices in S_Q2,n,2 are unitary. Indeed, if q is a linear combination of matrices in Q₂, q=x₁Q_i+x₂Q_j+x₃Q_k+x₄Q_l, and q is the conjugate matrix, q=x₁Q_i−x₂Q_j−x₃Q_k−x₄Q_l then q q=∥q∥²=(x₁²+x₂²+x₃²+x₄²)Q_i=kI. Therefore, all the elements in S_Q2n,2 are normalized by 1/√{square root over (k)} and they become unitary matrices.

Super-quaternions are constructed by Layers. Assume that the quaternion group Q_p has a matrix representation of m×m matrices. Let S_Qp,n,m be the collection of m×m matrices, which are all linear combinations of the matrices in Q_p, where the coefficients of the matrices are all integers and the sum of square roots of them is at most n. Then, S_Qp,n,m is written as a sequence of the layers

$\begin{array}{cc}{\bigcup }_{i=1}^nL_i,\mathrm{where}& \\ L_i=\left\{\sum _{s=1}^{2^p}\frac{x_sQ_s}{\sqrt{i}}\sum _{s=1}^{2^p}x_s^2=i,x_s\in Z\right\}.& \left(20\right)\end{array}$

When p=2 and m=2 then S_Q2,n,2 is a special case of this family. In this case, the set is the union of the layers L₁, . . . , L_n. L_i is the set of all linear combinations of the matrices in Q₂, where the coefficients x_i are integers that satisfy

x ₁ ² +x ₂ ² +x ₃ ² +x ₄ ²=1.  (21)

In this case, L₁ is Q₂ in Eq. 2 and it has eight elements. The calculation of the number of elements in each layer (which is the number of solutions to EQ. 21), shows that |L₂|=24, |₃|=32, |L₄|=24,etc.

By examining the layers L_i of the super quaternion, we observe that in some cases the same matrix element can exist in more than one layer. If (x₁, . . . , x_n) is in layer L_i, then (αx₁, αx₂, . . . , αx_n) must be in layer L_α2i, and since the elements are normalized, the matrices are equal. Therefore, for α>1 we have L_i⊂L_α2i. For example, the element Q_aεL₁ (the solution (1,0,0,0)), is equal to element 2Q_a/√{square root over (4)}εL₄ (the solution (2,0,0,0)). To eliminate the duplicate elements, we have to reduce these solutions in order to calculate correctly S_Q2,n,2 and its diversity product (gcd of all x_i is 1).

The Quaternion is a division algebra over R defined by D={a+bî+cĵ+d{circumflex over (k)}|a,b,c,dεR}, where the product of elements in D is defined by îĵ=−ĵî={circumflex over (k)}, î²=−1. Define f:D→M₂ such that f is a ring homomorphism from D to M₂, where M₂ is a set of 2×2 matrices over the complex field:

f(a+bî+cĵ+d{circumflex over (k)})=a·1+bQ_i+cQ_j+dQ_k.

If we define the M₂ norm as determinant, then it is easy to see that f is a homomoiphism. Since D is a division algebra every element in D is invertible, so D has no two sided ideals. This means that f is either an injection or the trivial zero transform (which it is obviously is not). Since f is an injection, all the matrices in its image are regular and furthermore the difference between any two matrices in M₂ is also regular. In other words, we construct a full diversity STBC code for two transmit elements.

4.2. Super Quaternion Diversity

In this section we present the diversity of the Super Quaternion structure. First we calculate the diversity of the commonly used orthogonal design for comparison. For each couple of constellation symbols S₁, S₂ we transmit the matrix

$C_{S_1S_2}=\left(\begin{array}{cc}{S}_1& S_2\\ -S_2^{*}& S_1^{*}\end{array}\right).$

The diversity is defined by

$\zeta =\frac{1}{2}\min \det \left(C_{S_1S_2}-C_{S_1^{\prime }S_2^{\prime }}\right){}^{\frac{1}{m}}=\frac{1}{2}\min \sqrt{{S_1-S_1^{\prime }}^2+{S_2-S_2^{\prime }}^2}$

where the min is over all the codewords C_S1S2 and C_S1′S2′. Without loss of generality, we can assume that S₁≠S′₁. In this case, we minimize the expression above by choosing S₂=S′₂. We can write ζ as:

$\begin{array}{cc}\begin{array}{c}\zeta =\frac{1}{2}\min S_1-S_1^{\prime }=\frac{1}{2}\min \left(S_1-S_1^{\prime }\right)S_1^{*}/S_1=\\ \frac{1}{2}\min {S_1}^2-S_1^{\prime }S_1^{*}/S_1.\\ \mathrm{Let}S_1=\frac{1}{\sqrt{2}}{}^{2k\pi i/n},S_1^{\prime }=\frac{1}{\sqrt{2}}{}^{2k^{\prime }\pi i/n}\end{array}& \left(22\right)\end{array}$

for n-PSK code (the normalization by 1/√{square root over (2)} aims to maintain transmit power of 1).

$\begin{array}{cc}\zeta =\frac{1}{2}\underset{k\ne k^{\prime }}{\min }\frac{1}{\sqrt{2}}1-{}^{2\left(k^{\prime }-k\right)\pi i/n}.& \left(23\right)\\ \begin{array}{c}{1-{}^{2\left(k^{\prime }-k\right)\pi i/n}}^2={1-{}^{2k^{″}\pi i/n}}^2\\ ={\left(1-\cos \left(2k^{″}\pi /n\right)\right)}^2+{\left(\sin \left(2k^{″}\pi /n\right)\right)}^2\\ =2\left(1-\cos \left(2k^{″}\pi /n\right)\right)=4{\sin }^2\left(k^{″}\pi /n\right).\end{array}& \left(24\right)\end{array}$

Substituting EQ. 24 in EQ. 23 we have

$\begin{array}{c}\zeta =\frac{1}{2}\underset{k\ne k^{\prime }}{\min }\frac{1}{\sqrt{2}}1-{}^{2\left(k^{\prime }-k\right)\pi i/n}=\frac{1}{2}\underset{k^{″}\ne 0}{\min }\frac{1}{\sqrt{2}}2\sin \left(k^{″}\pi /n\right)\\ =\frac{1}{\sqrt{2}}\underset{k^{″}\ne 0}{\min }\sin \left(k^{″}\pi /n\right)=\frac{1}{\sqrt{2}}\sin \left(\pi /n\right).\end{array}$

The following table summarizes the diversity of some of the quaternion structures: In this table, R is the transmission rate, L is the size of the constellation, M is the number of antennas and ζ is the diversity product.

					Group
	R	L	M	ζ	structure
	1.5	8	2	0.7071	Q2
	2.5	32	2	0.3827	SQ2,2,2
	3	64	2	0.3029	SQ2,3,2
	3.161	80	2	0.2588	SQ2,4,2
	3.5	128	2	0.1602	SQ2,5,2
	3.904	224	2	0.1602	SQ2,6,2
	4.085	288	2	0.1602	SQ2,7,2 = SQ2,8,2
	4.292	384	2	0.1374	SQ2,9,2
	4.522	528	2	0.0709	SQ2,10,2
	2.5	32	2	0.4082	L3
	2.29	24	2	0.5	L2
	2.79	48	2	0.3827	Q2 ∪ L2 ∪ L4

For comparison, the quaternion groups Q₂,Q₄ and Q₅ diversity is shown in the following table:

	R	L	M	ζ	Group structure
	1.5	8	2	0.7071	Quaternion group Q2
	2.5	32	2	0.1951	Quaternion group Q4
	3	64	2	0.0951	Quaternion group Q5

Generally, the diversity decreases as the rate increases, as can be seen from the table above.

4.2.1. Quaternion Group Example

In Q₂ there are eight code words (eight matrices):

$\begin{array}{c}\left(\begin{array}{cc}-1& 0\\ 0& -1\end{array}\right),\left(\begin{array}{cc}-i& 0\\ 0& i\end{array}\right),\left(\begin{array}{cc}0& -1\\ 1& 0\end{array}\right),\left(\begin{array}{cc}0& -i\\ -i& 0\end{array}\right),\left(\begin{array}{cc}0& i\\ i& 0\end{array}\right),\left(\begin{array}{cc}0& 1\\ -1& 0\end{array}\right),\\ \left(\begin{array}{cc}i& 0\\ 0& -i\end{array}\right),\left(\begin{array}{cc}1& 0\\ 0& 1\end{array}\right).\end{array}$

The rate for this code is ½ log₂ (8)=1.5 bits per channel use. The diversity of this code is ζ{Q₂}=0.707, which is equal to the diversity of Orthogonal Design [2] code for BPSK constellation that has a mere 1 bit per channel use rate.

FIG. 5 illustrates the Quaternion group constellation for rate 1.5.

4.2.2. Super Quaternion Set Example

The Super Quaternion set Q₂∪L₂∪L₄ contains 48 code words. The rate of this code is 2.7925 and the code diversity is 0.3827. By applying Orthogonal Design to 6-PSK we get merely a 0.3536 diversity for a lesser rate of 2.585. FIG. 6 illustrates the corresponding constellation.

5. FIXED-POINT-FREE GROUPS

This section elaborates on the groups G_m,r and J_m,r. We have chosen in the invention some of these groups, with the best performance and performed on them bit allocation. In some cases the groups were united with Super Quaternion codewords or some of the groups' members were removed in order to achieve optimal integral rate codes. Section 6 elaborates on methods to improve the performance of codes based on G_m,r groups.

5.1. The G_m,r Groups

Let

G _m,r = a,b:a ^m=1,bⁿ=a^t,bab⁻¹=a^r.

Since

(bab⁻¹)^r=ba^rb⁻¹=b(bab⁻¹)b⁻¹ba^rb⁻¹=bⁿ⁺¹ab⁻ⁿ⁻¹a^rn=bⁿab⁻ⁿ=a^rn=a^taa^−ta^rn=a.

We get the following restriction on r, n and m:

rⁿ≡(mod m).  (25)

Raising a^r to the power of t we get:

ba ^t b ⁻¹ =a ^rt bb ⁿ b−1=a^rtbⁿ=a^rta^t=a^rt.

Therefore,

$\begin{array}{cc}\left(r-1\right)t\equiv 1\left(\mathrm{mod}m\right)& \\ t=\frac{m}{gcd\left(m,r-1\right)}z,z\in Z.& \left(26\right)\end{array}$

To assure irreducible group we demand

gcd(n,t)=1.  (27)

We choose n such that n is the smallest integer that satisfies EQ. 25. We define z=1 in EQ. 26. The group is Fixed-Point-Free if all the prime divisors of n divide gcd(r−1,m) ([8]). The group G_mr has mn elements because it contains the subgroup <a> of order m and n distinct cosets:

G _m,r = a ∪ a b∪ . . . ∪ a b ⁿ⁻¹.

There is an embedding

L_G:G_m,r→K_G

of G_m,r in the division algebra K_G defined as follows. Let K be the cyclotomic of by an m-th root of 1, ζ, and K_G the cyclic algebra defined by the relations bⁿ=ζ_m^t=γ, bζ_mb⁻¹=σ(ζ_m)bb⁻¹=ζ_m^r.

L_G is defined by the formulas

L _G(a)=ζ_m, L_G(b)=b.

We take ζ_m=e^2πi/m.

A matrix representation of K_G:

Π:K_G→M_n( )

where M_n is the algebra of n×n matrices over is given by:

$\begin{array}{cc}\prod \left({\xi }_m\right)=\left(\begin{array}{ccccc}{\xi }_m& 0& 0& \dots & 0\\ 0& {\xi }_m^r& 0& \dots & 0\\ 0& 0& {\xi }_m^{r^2}& \dots & 0\\ ⋮& ⋮& ⋮& ⋰& ⋮\\ 0& 0& 0& \dots & {\xi }_m^{r^{n-1}}\end{array}\right),\prod \left(b\right)=\left(\begin{array}{ccccc}0& 1& 0& \dots & 0\\ 0& 0& 1& \dots & 0\\ ⋮& ⋮& ⋮& ⋰& ⋮\\ 0& 0& 0& \dots & 1\\ {\xi }_m^t& 0& 0& \dots & 0\end{array}\right).& \left(28\right)\end{array}$

Explicitly, the element a₀+a₁b+a₂b²+ . . . +a_n−1bⁿ⁻¹ maps to

$\left(\begin{array}{ccccc}{a}_0& a_1& a_2& \dots & a_{n-1}\\ \gamma \sigma \left(a_{n-1}\right)& \sigma \left(a_0\right)& \sigma \left(a_1\right)& \dots & \sigma \left(a_{n-2}\right)\\ \gamma {\sigma }^2\left(a_{n-2}\right)& \gamma {\sigma }^2\left(a_{n-1}\right)& {\sigma }^2\left(a_0\right)& \dots & {\sigma }^2\left(a_{n-3}\right)\\ \dots & \dots & \dots & \dots & \dots \\ \gamma {\sigma }^{n-1}\left(a_1\right)& \gamma {\sigma }^{n-1}\left(a_2\right)& \gamma {\sigma }^{n-1}\left(a_3\right)& \dots & {\sigma }^{n-1}\left(a_{n-2}\right)\end{array}\right)\hspace{1em}$

The element ‘1’ is transformed to the identity matrix I_n×n. All the matrices in this representation are unitary and therefore this code can be implemented for unknown channels as well as for known channels.

5.2. The J_m,r Groups

We present here a specific fixed point group with exceedingly good performance which are due to Amitsur ([13])

J _m,r =SL ₂(F₅)×G_m,r  (29)

where m,r are as in section 5.1, gcd(mn,120)=1, and SL₂(F₅) is the group of 2×2-matrices over F₅ with determinant 1. SL₂(F₅) has a presentation

SL ₂(F₅)=u,γ|μ²=γ³=(μγ)⁵,μ⁴=1  (30)

The order of J_m,r is 120mn.

We define the following matrix representation:

Π:J_m,r→M_2n(F₅)

where M_2n(F₅) is the set of 2n×2n-matrices over F₅.

$\prod \left(\mu \right)=P=P_0\otimes I_n,P_0=\frac{1}{\sqrt{5}}\left(\begin{array}{cc}{\eta }^2-{\eta }^3& \eta -{\eta }^4\\ \eta -{\eta }^4& {\eta }^3-{\eta }^2\end{array}\right),\prod \left(\gamma \right)=Q=Q_0\otimes I_n,Q_0=\frac{1}{\sqrt{5}}\left(\begin{array}{cc}\eta -{\eta }^2& {\eta }^2-1\\ 1-{\eta }^3& {\eta }^4-{\eta }^3\end{array}\right)$

where η=ζ₅=e^2πi/5, I_n is the n×n and denotes the Kronecker product.

$\prod \left(a\right)=A=I_2\otimes \left(\begin{array}{ccccc}{\xi }_m& 0& 0& \dots & 0\\ 0& {\xi }_m^r& 0& \dots & 0\\ 0& 0& {\xi }_m^{r^2}& \dots & 0\\ ⋮& ⋮& ⋮& ⋰& ⋮\\ 0& 0& 0& \dots & {\xi }_m^{r^{n-1}}\end{array}\right)$ $\prod \left(b\right)=B=I_2\otimes \left(\begin{array}{ccccc}0& 1& 0& \dots & 0\\ 0& 0& 1& \dots & 0\\ ⋮& ⋮& ⋮& ⋰& ⋮\\ 0& 0& 0& \dots & 1\\ {\xi }_m^t& 0& 0& \dots & 0\end{array}\right)$

where ζ_m is defined in section 5.1. The constellation consists of the matrices

A^sB^k(PQ)^jX, s=0, . . . , m−1, k=0, . . . , n−1, j=0, . . . , 9

• and Xε{I_2n,P,Q,QP,QPQ,QPQP,QPQ², QPQPQ, QPQPQ², QPQPQ²P, QPQPQ²PQ, QPQPQ²PQP}.The smallest group of type J_m,r is J_1,1, which is isomorphic to SL₂(F₅), and has 120 elements (3.5 rate). This group is given by the matrices (PQ)^jX. This group has diversity of 0.3090 and exceeding performance in comparison to all other methodologies shown in this invention for rate 3.5 and 2-transmit diversity as shown in FIG. 13. In order to enable bit-allocation we extended the J_1,1 group to a 128-element set using Quaternion matrices from layers 2 and 3 (EQ. 20).

6. COSETS OF FIXED-POINT-FREE GROUPS—NEW STBC

The basic idea in using cosets is that if we have a unitary matrix that has good diversity with a specific group, then this diversity will be preserved when taking the union of the group and the coset determined by this element. Indeed, if G is a group of unitary matrices, and h is a unitary matrix not in G, we can take the coset determined by h, i.e. all the elements of the form hg, where g E G. If |h−g|>d,∀gεG, and the diversity of G is at least d, then we claim that this will be the diversity of the union. This is true since

|hg₁−hg₂|=|h∥g₁−g₂|=|g₁−g₂|>d

and

|hg₁−g₂|=|(h−g₂g₁⁻¹)g₁|=|h−g₂g₁⁻¹∥g₁|>d.

We can now repeat this process, choosing a unitary matrix that has good diversity with the union already chosen. In fact, there are cases where it is advantageous to choose a pair of matrices simultaneously and add the cosets of both matrices. In this section we construct large sets of “unitary” elements in division algebras that are finite dimensional over Q. The construction will occur in “cyclic algebras” which we now describe. In Particular we will elaborate on the implementations of this method on the groups G_mr.

6.1. Cyclic algebras

Let K/k be a cyclic Galois extension of dimension n, so that its Galois group is cyclic of order n. Let σ be a generator of the Galois group. Assume that 0≠γεk. The cyclic algebra associated with this data is defined, as a left K vector space, as

K⊕Kb⊕ . . . ⊕Kbⁿ⁻¹.

The multiplication is be defined by the following equations

bⁿ=γ

ba=σ(a)b ∀aεK

We will denote such an algebra as (K/k, γ).

Let m,r be relatively prime integers, and let n be the order of r in the multiplicative group (Z/(m))*. Let s=(r−1,m) and t=m/s. Suppose that (n,t)=1 and n|s. We can now define a central simple algebra over Q. This algebra is constructed by taking the cyclotomic extension, K, generated by the roots of unity of order m. The Galois group of this extension is the multiplicative group (Z/(m)), and r defines a cyclic subgroup of order n in the Galois group, whose generator is an automorphism, σ, which raises roots of unity to the r^th power. The center of our algebra will be the fixed subfield of this automorphism, k, and σ is a generator of the Galois group of the extension K/k. Let ζ be a primitive root of unity of order m, and denote γ=ζ^t. The algebra is defined as the cyclic algebra A_m,r=(K/k,γ). For certain values of m,r, A_m,r is a division algebra, for instance (m,r)=(21,4). The precise conditions are quite complicated, and are spelled out completely by Amitsur [13].

There is a natural embedding of this algebra in M_n( ), which comes from the regular representation of the algebra as a left vector space of dimension n over K. Explicitly, the element a₀+a₁b+a₂b²+ . . . +a_n−1bⁿ⁻¹ maps to

$\left(\begin{array}{ccccc}{a}_0& a_1& a_2& \dots & a_{n-1}\\ \gamma \sigma \left(a_{n-1}\right)& \sigma \left(a_0\right)& \sigma \left(a_1\right)& \dots & \sigma \left(a_{n-2}\right)\\ \gamma {\sigma }^2\left(a_{n-2}\right)& \gamma {\sigma }^2\left(a_{n-1}\right)& {\sigma }^2\left(a_0\right)& \dots & {\sigma }^2\left(a_{n-3}\right)\\ ⋮& ⋮& ⋮& ⋰& ⋮\\ \gamma {\sigma }^{n-1}\left(a_1\right)& \gamma {\sigma }^{n-1}\left(a_2\right)& \gamma {\sigma }^{n-1}\left(a_3\right)& \dots & {\sigma }^{n-1}\left(a_0\right)\end{array}\right)\hspace{1em}$

A quick check shows that all the fixed-point-free groups of type G_m,r can be embedded into A_m,r so we will be looking for “unitary” elements in

6.2. Finding Unitary Elements in Cyclic Algebras

In order to find “unitary” elements in A_m,r we define an anti-automorphism, τ, of the algebra, which corresponds to taking the conjugate transpose of a matrix. The “unitary” elements will be those such that τ(x)=x⁻¹. The field k will be invariant under τ. Thus, it is enough to define τ on a primitive root of unity, and we take τ(ζ)=ζ⁻¹. This is simply the restriction of complex conjugation to K under a fixed embedding of K in C. It remains to define τ on b, and since b should be unitary, we must have τ(b)=b⁻¹=γ⁻¹bⁿ⁻¹. From the requirement that τ be an anti-automorphism, this defines τ completely, and it is easy to see that it is well defined. Note that τ²=id.

We are now in a position to find many unitary n X n matrices that have positive diversity, since all the elements we find will be in a division algebra. In fact, we will need to divide by square roots. If n is odd, then adding a square root cannot split the algebra, hence we will still have a division algebra.

When looking for unitary elements, we are looking for

$x=\sum _{i=0}^{n-1}a_ib^i$

Note that

$\tau \left(x\right)=\tau \left(a_0\right)+\sum _{i=1}^{n-1}{\gamma }^{-1}{\sigma }^i\left(\tau \left(a_{n-i}\right)\right)b^i.$

Now, we require xτ(x)=1. If z=xτ(x) then let

$z=\sum _{i=0}^{n-1}{\alpha }_ib^i.$

It is easy to see that τ(z)=z so we get α₀=τ(a₀) and α_i=γ⁻¹σ_i(τ(α_n−i)). For instance, if n=3, the equation xτ(x)=1 is really just two (and not three) equations, because if α₁=0 then automatically α₂=0.

Continuing with the case n=3, let us consider the case where a₀, a₁ are known, and are in our extended algebra. In this case we get

xτ(x)=(a₀+a₁b+a₂b²)(τ(a₀)+γ⁻¹σ(τ(a₂))b+γ⁻¹σ²(τ(a₁))b²)=(a₀τ(a₀)+a₁τ(a₁)+a₂τ(a₂))+(γ⁻¹a₀σ(τ(a₂))+a₁σ(τ(a₀))+γ⁻¹σ(τ(a₁)a₂))b+a₂b²

so there are really only two equations, the second of which is of the form

ασ(τ(a₂))+βa₂=δ

where α=γ⁻¹a₀, β=γ⁻¹σ(τ(a₁)), δ=−a₁σ(τ(a₀)), and in particular α,β,δεK.

Since K is a vector space over Q of dimension φ(m), we get φ(m) linear equations in the φ(m) coordinates of a₂ as an element in K.

Thus, given a₀, a₁εK, we check if there are solutions to the linear set of equations. For each solution, a₂, we calculate

a ₀τ(a₀)+a₁τ(a₁)+a₂τ(a₂)=s(a₀,a₁,a₂)εK.

We divide all three values by √{square root over (s(a₀,a₁,a₂))}, to get a unitary matrix. Note, that to calculate σ(√{square root over (s)}) we can simply take √{square root over (σ(s))}.

If a₀, a₁ are rational, it can be seen that if a₀²≠a₁² then there is a solution. Indeed, if we set a₂=xγ+yγ⁻¹, we have σ(a₂)=a₂ and τ(a₂)=yγ+xγ⁻¹. The equation becomes

γ⁻¹a₀(yγ+xγ⁻¹)+γ⁻¹a₁(xγ+yγ⁻¹)=−a₀a₁.

Thus, we have in fact two equations a₀y+a₁x=−a₀a₁ and a₁y+a₀x=0. There is one

$x=\frac{a_0a_1^2}{a_0^2-a_1^2},y=-\frac{a_0^2a_1}{a_0^2-a_1^2}.$

6.3. Finding Elements for Cosets

The following refers to trying to find a good coset extension of some G_m,r group, which has n=3 (i.e. describes 3×3 matrices). Using the terminology above, we know that there is an infinite number of unitary matrices we can try, simply by taking a₀, a₁ to be rational. In fact, if we take a₀=1 and a₁ very small, we will get a unitary matrix that is very close to being the identity. This gives us the ability to make small “changes” to elements of the algebra.

We can now describe two different methods for trying to find cosets. In the first method, we construct a number of elements of the algebra in the way described above. We then simply take random multiplications of these elements. For each new element we get we check the diversity against the set of matrices we have so far. It is easy to see that this will be the diversity of the entire coset. We do this many times, and choose the coset that has the best diversity.

The second method is to construct a set of matrices that are close to the identity. One way to do this is to take a₁=1, and for each basis element of K over Q and each of the three elements of the first row of the matrix, we add a small rational multiple of this basis element. We can decrease this multiple as we go along, so that we make smaller and smaller changes. If we have a set of matrices, and a trial matrix, we try changing it “slightly” in all the directions, and take the best one. If we can no longer improve, we decrease the multiple and try again. We can also try to add two or more cosets at a time, and try changes in one or more of the matrices. So far, when attempting to add two cosets, the best method seems to be taking two additional elements and trying to change both of them slowly.

The following refers to trying to find a coset extension of some G_m,r group, which has n=3 (i.e. describes 3×3 matrices). This fixed-point-free group resides in a cyclic algebra, which has the following structure. Take F to be the cyclotomic extension of the rationals by the roots of unity of order m. Let ζ be a primitive root of unity of order m, and we can assume that r̂3=1 modulo m (n=3). Thus, we can also assume that 3 divides m, and we set s=gcd(m,r−1), t=m/s. We now define the cyclic algebra A over F by adjoining an element b such that b̂3=ζ̂t. The fixed-point-free group will be the group generated by ζ and b, and will be of order 3m. When we want to take cosets of the group, we would like to find some more unitary elements in the algebra (since when the algebra is a division ring, we know that the diversity will remain non-zero). In order to do so, we simply check what will have a unitary image. Any element of the algebra will be defined by 3 elements in F, a₁, a₂, a₃. They define the matrix

$\left(\begin{array}{ccc}{a}_1& a_2& a_3\\ \gamma \sigma \left(a_3\right)& \sigma \left(a_1\right)& \sigma \left(a_2\right)\\ \gamma {\sigma }^2\left(a_2\right)& \gamma {\sigma }^2\left(a_3\right)& {\sigma }^2\left(a_1\right)\end{array}\right)\hspace{1em}$

where γ=ζ^t and σ is the Galois automorphism of F over the rationals. For such a matrix to be unitary we need that each row has absolute value (as complex numbers) 1, and that each pair of rows are orthogonal, in the sense that the dot product of row by the complex conjugate of the other row is zero. A simple calculation shows that in the matrices given above, it is enough to check the absolute value of the first row, and whether the first two rows are orthogonal (note that this is true only for 3×3 matrices. If n is larger, there are conditions that need to be met). Note, that if we have a matrix where the two first rows are orthogonal, but the absolute value is k, then we can always divide the first row by the square root of k, the second row by the square root of σ(k), and the last row by the square root of σ²(k). Note that adjoining these square roots to F cannot split the algebra, since if is of odd degree over F. Suppose now that we give a₁ and a₂ some set values. The orthogonality condition becomes (note that σ and the complex conjugate commute)

ya ₁σ( a₃)+σ( a₂)a₃=−a₂σ( a₁)

Since a₁ and a₂ are known, we get linear equations in a₃ and σ( a₃). However, if we consider a₃ as a rational linear of the basis of F over the rationals, then both a₃ and σ( a₃) are linear combinations of the coefficients. F is of dimension φ(m) over the rationals, and we get a set of φ(m) equations in φ(m) variables. If the equations are of full rank, then there will be exactly one solution. If they are not of full rank we might get no solutions. In the case where a₁ and a₂ are both rational, it is easily seen that if a₁²≠a₂² then there is exactly one solution. This already shows us that there is an infinite number of unitary elements we can use, and in fact, if we take a₁=1 and a₂ very small, we will get a unitary matrix that is very close to being the identity. This gives us the ability to make small “changes” to elements of the algebra.

We can now describe two different methods for trying to find cosets. In the first method, we construct a number of elements of the algebra in the way described above. We then simply take random multiplications of these elements. For each new element we get we check the diversity against the set of matrices we have so far It is easy to see that this will be the diversity of the entire coset. We do this many times, and choose the coset that has the best diversity.

The second method is to construct a set of matrices that are close to the identity. One way to do this is to take a₁=1, and for each basis element of F over the rationals and each of the three elements of the first row of the matrix, we add a small multiple of this basis element. We can decrease this multiple as we go along, so that we make smaller and smaller changes. If we have a set of matrices, and a trial matrix, we try changing it “slightly” in all the directions, and take the best one. If we can no longer improve, we decrease the multiple and try again. We can also try to add two or more cosets at a time, and try changes in one or more of the matrices. So far, when needing three cosets in total (including the original group) the best method seems to be taking two additional elements and trying to change both of them slowly.

6.4. Results of Coset Search

When attempting to find a good set of 512 3×3 unitary matrices, using the G_m,r groups, the best diversity we can get is 0.184, for m=186,r=25 (there are actually 558 matrices, but for practical purposes it is often best to take a power of 2 number of matrices). However, when taking m=63,r=37 and adjoining two additional cosets (using the second method mentioned above, adding both cosets at the same time), we achieved a diversity of 0.224. When running simulations of error rates compared to SNR, we see that around an error rate of 10⁻², the new set has an SNR that is better by almost 1 db than the best G_m,r group.

7. BIT ALLOCATION

An important issue in any code constellation (higher than 2) is the bit allocation. Bit allocation is the look-up table that assigns bits to symbols. In codes with integral rate, i.e. codes that have 2ⁿ code words (i.e. the code rate is an integer number), the bit allocation is simple. In MIMO channels, if the rate is N (m is the transmit diversity), we can still allocate bits for each symbol using look-up table, because each matrix includes m channel uses. In order to take better advantage of the rate, for other rates, we have to allocate bits for sequences of symbols. This allocation may result in a very high complexity. In this work we will limit ourselves to an integral rate bit allocation.

In cases we would use codes with not integral rates, the allocation is made by maximizing the diversity over the maximal subgroup of code words that has an integral rate. For example the Super Quaternion structure S_Q2,8,2 has 288 code words. By selecting 256 code words we create an integral rate(4) code with improved performance.

Consider a code with 2ⁿ code words and rate n/m. First we calculate the BER from the FRAME ERROR if we choose a random bit allocation. Assume we have mistakenly reconstructed in the receiver a code word e, while c was transmitted. Since the bit allocation was randomly chosen, we can calculate the average number of bits interpreted wrongly by

$\begin{array}{c}E\left\{cXORec\ne e\right\}=E\left\{cXORecXORe\ge 1\right\}\\ =\frac{2^nn}{2\left(2^n-1\right)}=\frac{n}{2}+\frac{n}{2\left(2^n-1\right)}.\end{array}$

It follows that for random bit allocation the ratio between BER and FER(Frame Error Rate) is

$\frac{1}{2}\left(1+\frac{1}{2^n-1}\right).$

However in channels with small noise most of the errors occur due to reconstruction of symbols, which are close to the transmitted original. If the bit allocation is optimally made, close symbols will differ by minimal number of bits (merely a single bit if possible) and at high SNR the ratio between BER and FER will be approximately 1/n. Thus, we can reach the same BER as in random bit allocation with lower SNR (adjusted for FER higher by

$\frac{n}{2}\left(1+\frac{1}{2^n-1}\right)).$

FIG. 7 shows an example for Gray-code allocation for 8PSK constellation, which preserves a single bit difference between neighboring symbols. The significance of bit allocation can be seen in FIGS. 9 and 12.

8. RATE BOUND FOR 2-TRANSMIT DIVERSITY AND EXHAUSTIVE SEARCH

This section presents general bounds for 2-transmit diversity both for the case of orthogonal design and the specific case of unitary design.

8.1. Isometry to R⁴

Let M₂ be a set of orthogonal matrices of the form

$m_2=\left(\begin{array}{cc}{S}_1& S_2\\ -S_2^{*}& S_1^{*}\end{array}\right)$

for each m₂±M₂ where S₁, S₂ are some complex symbols.

We define the following isometry

g:M₂→R⁴  (31)

Let

$m_2=\left(\begin{array}{cc}{S}_1& S_2\\ -S_2^{*}& S_1^{*}\end{array}\right)$

where S₁=a+bi and S₂=c+di. Then g(m₂)=(a,b,c,d). It is easy to see that g is an isometry if we define the distance between two matrices m₂, m₂′ in M₂ to be

$\begin{array}{cc}d\left(m_2,m_2^{\prime }\right)=\frac{1}{2}\sqrt{\det \left(m_2-m_2^{\prime }\right)}& \left(32\right)\end{array}$

and the distance between two vectors in R⁴ to be half the Euclidian distance:

$\begin{array}{cc}d\left(\left(a,b,c,d\right),\left(a^{\prime },b^{\prime },c^{\prime },d^{\prime }\right)\right)=\frac{1}{2}\sqrt{\begin{array}{c}{\left(a-a^{\prime }\right)}^2+{\left(b-b^{\prime }\right)}^2+\\ {\left(c-c^{\prime }\right)}^2+{\left(d-d^{\prime }\right)}^2\end{array}}.& \left(33\right)\end{array}$

The minimal distance among the matrices in M₂, according to EQ. 32 is the diversity of M₂ by definition.

8.2. Orthogonal Design

Assume we have a set of n vectors {(a_i,b_i,c_i,d_i)}_i=1ⁿ with minimal distance (diversity) ζ, and maximal norm (distance from (0,0,0,0)) A. What is the relation among the parameters n, ζ and A?

Every vector {(a_i,b_i,c_i,d_i)} is surrounded by a 4-dimensional sphere with radius ζ/2 that does not have other vectors. The volume of this sphere is

$\begin{array}{c}\int \int {\int }_{{\int }_{w^2+x^2+y^2+z^2\le {\zeta }^2}}wxyz={\int }_{R^2+z^2={\zeta }^2}\frac{4\pi }{3}R^3z=\\ {\int }_{z=-\zeta }^{+\zeta }\frac{4\pi }{3}{\left(\zeta -z^2\right)}^{\frac{1}{2}}z\end{array}$

Define zζ cos θ, dz−ζ sin θdθ. Then

$\begin{array}{cc}\begin{array}{c}\begin{array}{c}{\int }_{z=-\zeta }^{+\zeta }\frac{4\pi }{3}{\left(\zeta -z^2\right)}^{\frac{3}{2}}z={\int }_{-\pi }^0\frac{4\pi }{3}{\zeta }^3{\sin }^3\theta \left(-\zeta \sin \theta \right)\theta =\\ {\int }_{-\pi }^0\frac{4\pi }{3}{\zeta }^4{\sin }^4\theta \theta ={\int }_{-\pi }^0\frac{4\pi }{3}{\zeta }^4\left({\sin }^2\theta -\frac{1}{4}{\sin }^22\theta \right)\theta =\end{array}\\ \frac{4\pi }{3}{\zeta }^4\left(\frac{1}{2}\pi -\frac{1}{8}\pi \right)=\frac{{\pi }^2}{2}{\zeta }^4.\end{array}& \left(34\right)\end{array}$

Since all the vectors have at most norm A, all the spheres that are defined in EQ. 34 are confined within the sphere with radius A+ζ/2, which has the volume

$\begin{array}{cc}\frac{{\pi }^2}{2}{\left(2A+\zeta \right)}^4.& \left(35\right)\end{array}$

It follows that

$n\le \frac{\frac{{\pi }^2}{2}{\left(2A+\zeta \right)}^4}{\frac{{\pi }^2}{2}{\zeta }^4}={\left(\frac{2A+\zeta }{\zeta }\right)}^4.$

Given a maximal transmit power of a matrix in M₂, i.e. max{det(m₂)|m₂εM₂}=P, and a minimal diversity ζ, the maximal number of matrices in M₂ is

$\begin{array}{cc}n\le {\left(\frac{\sqrt{P}+\zeta }{\zeta }\right)}^4& \left(36\right)\end{array}$

and the maximal rate is

R _max≦2(log₂(√{square root over (P)}+ζ)−log₂(ζ)).  (37)

8.3. Unitary Design

For unitary matrices a tighter bound can be found. Since unitary matrices have determinant 1, the isometry of a unitary set of matrices to R, is equivalent to placing vectors on the envelope of a 4-dimensional sphere, whose volume is given by EQ. 34. The ‘area’ of this envelope is

$\begin{array}{cc}\frac{}{x}{\left(\frac{{\pi }^2}{2}x^4\right)}_{x=1}=2{\pi }^2& \left(38\right)\end{array}$

In order to calculate the free space around each vector on the envelope of the four dimensional sphere we change variables:

$\begin{array}{cc}\begin{array}{c}x\stackrel{\Delta }{=}R\cos \left(\gamma \right)\cos \left(\theta \right)\cos \left(\varphi \right)\\ y\stackrel{\Delta }{=}R\cos \left(\gamma \right)\cos \left(\theta \right)\sin \left(\varphi \right)\\ z\stackrel{\Delta }{=}R\cos \left(\gamma \right)\sin \left(\theta \right)\\ w\stackrel{\Delta }{=}R\sin \left(\gamma \right).\end{array}& \left(39\right)\end{array}$

The absolute value of the Jacobian of these variables is:

$\begin{array}{cc}\begin{array}{cccc}\cos \left(\gamma \right)\cos \left(\theta \right)\cos \left(\varphi \right)& \cos \left(\gamma \right)\cos \left(\theta \right)\sin \left(\varphi \right)& \cos \left(\gamma \right)\sin \left(\theta \right)& \sin \left(\gamma \right)\\ -R\sin \left(\gamma \right)\cos \left(\theta \right)\cos \left(\varphi \right)& -R\sin \left(\gamma \right)\cos \left(\theta \right)\sin \left(\varphi \right)& -R\sin \left(\gamma \right)\sin \left(\theta \right)& R\cos \left(\gamma \right)\\ -R\cos \left(\gamma \right)\sin \left(\theta \right)\cos \left(\varphi \right)& -R\cos \left(\gamma \right)\sin \left(\theta \right)\sin \left(\varphi \right)& R\cos \left(\gamma \right)\cos \left(\theta \right)& 0\\ R\cos \left(\gamma \right)\cos \left(\theta \right)\sin \left(\varphi \right)& -R\cos \left(\gamma \right)\cos \left(\theta \right)\cos \left(\varphi \right)& 0& 0\end{array}=R^3{\cos }^2\left(\gamma \right)\cos \left(\theta \right)& \left(40\right)\end{array}$

Without loss of generality we calculate the free space around the vector {right arrow over (r)}₀=(w₀,x₀,y₀,z₀), (equivalently defined by R₀,γ₀,θ₀ and φ₀) on the w axis, i.e.:

w₀=1, x₀=y₀=z₀=0

or in our new coordinates:

R₀=1, γ₀=π/2.

For diversity ζwe calculate the free three-dimensional area around {right arrow over (r)}₀ by integrating over the envelope of the sphere and over the vectors within distance smaller than ζ/2 (by saying distance we mean the definition in Eq. 3). The vector {right arrow over (r)} on the sphere satisfies:

$\begin{array}{c}{(}_{\stackrel{->}{r}},{\stackrel{->}{r}}_0)=\frac{1}{2}\sqrt{x^2+y^2+z^2+{\left(w-1\right)}^2}\\ =\frac{1}{2}\sqrt{2-2w}=\frac{1}{2}\sqrt{2-2\sin \left(\gamma \right)}.\end{array}$

The free ‘area’ is therefore confined to

${(}_{\stackrel{->}{r}},{\stackrel{->}{r}}_0)=\frac{1}{2}\sqrt{2-2\sin \left(\gamma \right)}\le \zeta /2$ $\sin \left(\gamma \right)\ge 1-\frac{1}{2}{\zeta }^2$ $\gamma \ge \arcsin \left(1-\frac{{\zeta }^2}{2}\right).$

Now, we can calculate the free ‘area’ around {right arrow over (r)}₀:

$\int {\int }_{\int {(}_{\stackrel{->}{r}},{\stackrel{->}{r}}_0)\le \zeta /2}R^3{\cos }^2\left(\gamma \right)\cos \left(\theta \right)\gamma \theta \varphi$ $\begin{array}{c}=\int {\int }_{\int \gamma \ge \arcsin \left(1-\frac{{\zeta }^2}{2}\right)}{\cos }^2\left(\gamma \right)\cos \left(\theta \right)\gamma \theta \varphi \\ ={\int }_0^{2\pi }\varphi {\int }_{\frac{-pi}{2}}^{\frac{pi}{2}}\cos \left(\theta \right)\theta {\int }_{\arcsin \left(1-\frac{{\zeta }^2}{2}\right)}^{\frac{pi}{2}}{\cos }^2\left(\gamma \right)\gamma \\ =2\pi ·\sin \left(\theta \right){}_{-\pi /2}^{\pi /2}·{\int }_{\arcsin \left(1-\frac{{\zeta }^2}{2}\right)}^{\frac{pi}{2}}(\frac{1}{2}\left(1+\cos \left(2\gamma \right)\right)\gamma \\ =2\pi \left[\frac{\pi }{2}-\arcsin \left(1-\frac{{\zeta }^2}{2}\right)-\frac{1}{2}\sin \left(2\arcsin \left(1-\frac{{\zeta }^2}{2}\right)\right)\right].\end{array}$

Using the identity:

sin(2arc sin x)=2 sin(arc sin x)cos(arc sin x)=2x√{square root over (1−x²)}

we get that the free ‘area’ is:

$\begin{array}{cc}\int \int \int {(}_{\stackrel{->}{r}},{\stackrel{->}{r}}_0)\le \zeta /2R^3{\cos }^2\left(\gamma \right)\cos \left(\theta \right)\gamma \theta \varphi =2\pi \left[\frac{\pi }{2}-\arcsin \left(1-\frac{{\zeta }^2}{2}\right)-\zeta \left(1-\frac{{\zeta }^2}{2}\right)\sqrt{1-\frac{{\zeta }^2}{4}}\right].& \left(41\right)\end{array}$

The size of a unitary constellation with diversity ζ according to EQS. 41 is smaller than

$\begin{array}{cc}n\le \frac{\pi }{\frac{\pi }{2}-\arcsin \left(1-\frac{{\zeta }^2}{2}\right)-\zeta \left(1-\frac{{\zeta }^2}{2}\right)\sqrt{1-\frac{{\zeta }^2}{4}}}.& \left(42\right)\end{array}$

and the maximal rate is

$\begin{array}{cc}{R}_{\max }\le \frac{1}{2}\left(\begin{array}{c}{\log }_2\left(\pi \right)-{\log }_2(\frac{\pi }{2}-\arcsin \left(1-\frac{{\zeta }^2}{2}\right)-\\ \zeta \left(1-\frac{{\zeta }^2}{2}\right)\sqrt{1-\frac{{\zeta }^2}{4})}\end{array}\right).& \left(43\right)\end{array}$

EXAMPLES

The group J_1,1 (given in EQ. 29) has the diversity=0.309 and rate R=3.45, while the bound on a unitary set with such diversity is R_max≦3.66 (EQ. 43).

The Super Quaternion group Q₂ωL₂ωL₄ has diversity ζ=0.3827 and rate R=2.79, while the bound on the rate of a set with diversity ζ=0.3827 is R_max≦3.2.

The Super Quaternion set S_Q2,8,2 has diversity ζ=0.1602 and rate R=4, while the bound on the rate of a set with such diversity is R_max≦5.

8.4. Exhaustive Search

The bounds calculated above gives us an insight on the characteristics of good codes. The isometry g of good orthogonal codes will result in uniformly distributed vectors in a 4-dimensional sphere, while the isometry of good unitary codes will result in vectors that are uniformly distributed on the envelope of a 4-dimensional sphere with radius 1.

One way to achieve uniform distribution of unitary codes is to generate the code by exhaustive search over γ, θ, and φ (EQ. 39). The exhaustive search we present here is based on the following set in R⁴

V={(x,y,z,w)|x=cos(γ)cos(θ)cos(φ),y=cos(γ)cos(θ)sin(φ),z=cos(γ)sin(θ),w=sin(γ),γ=kC+γ₀,θ=lC/cos(γ)+θ₀(γ),φ=mC/cos(γ)/cos(θ)+φ₀(γ,θ),k,l,mεN,0<C<π}  (44)

The search is done over C, which changes monotonically with the diversity and over γ₀, θ₀ (γ), and φ₀(γ,θ). Each vector set is isomorphic to a unitary code according to EQ. 31. Results of exhaustive search are shown in FIGS. 14 and 15. The performance of the search are typically close to the Super Quaternion sets, and outperforms the Orthogonal Design at high rates.

9. UNKNOWN CHANNELS

Symbols in unknown SISO channels are transmitted differentially to each other. If at time t we intend to send the symbol S_t and the previous transmitted symbol was X_t−1, we transmit X_t−1·S_t, i.e. at time t we transmit

$X_t=\prod _{k=0}^tS_k.$

Assume that the channel is quasi-static (changes much slower than the symbol rate), then the symbols can be differentially decoded over unknown channel. This transmit algorithm is only effective for phase and frequency modulations since amplitude modulations are more difficultly separated from channel fading.

9.1. Encoding Algorithm

Due to the unitarity of the code used (shown in section 4) it is also applicable for an unknown channel. All the unitary codes can be transmitted over unknown channels. We use the SISO differential decoding principles for MIMO unknown channels in the following way: Instead of multiplying symbols we multiply matrices. If at time t−1 we transmitted the matrix X_t−1 and the next matrix we want to send is S_t then we would transmit at time t the matrix

X _t =S _t ·X _t−1.  (45)

It is important to notice that for n transmitters, each matrix contains n time samples, i.e. when we refer to transition at time t we actually refer to the n time samples

nk, nk+1, . . . , (n+1)k−1

where nk=t.

We assume m receiver antennas. Define the channel response at time t as

$R_t=R=\left(\begin{array}{cccc}{\alpha }_{1,1}& {\alpha }_{1,2}& & {\alpha }_{1,m}\\ {\alpha }_{2,1}& {\alpha }_{2,2}& & {\alpha }_{2,m}\\ ⋮& ⋮& ⋰& ⋮\\ {\alpha }_{n,1}& {\alpha }_{n,2}& & {\alpha }_{n,m}\end{array}\right)$

where α_i,j is the response from transmitter i to receiver j. We removed the time dependency (n) due to the quasi-static channel assumption. Let the channel noise be

${\Upsilon }_t=\left(\begin{array}{cccc}{\upsilon }_t^{1,1}& {\upsilon }_t^{1,2}& & {\upsilon }_t^{1,m}\\ {\upsilon }_t^{2,1}& {\upsilon }_t^{2,2}& & {\upsilon }_t^{2,m}\\ ⋮& ⋮& ⋰& ⋮\\ {\upsilon }_t^{n,1}& {\upsilon }_t^{n,2}& & {\upsilon }_t^{n,m}\end{array}\right)$

where ν_t^i,j is the noise at antenna j at time t+i. The received symbols at time t are

$Y_t=\left(\begin{array}{cccc}{y}_t^{1,1}& y_t^{1,2}& & y_t^{1,m}\\ y_t^{2,1}& y_t^{2,2}& & y_t^{2,m}\\ ⋮& ⋮& ⋰& ⋮\\ y_t^{n,1}& y_t^{n,1}& & y_t^{n,m}\end{array}\right)$

where y_t^i,j is the received symbol at antenna j at time t+i.

Y _t =X _t R+Υ _t.

9.2. Decoding Algorithm

The decoding is performed by estimating every codeword S_t according to the last two received matrices Y_t and Y_t−1. The decoding algorithm is given by

$\begin{array}{cc}{S}_n=\mathrm{SYMBOL}\left[\underset{k}{\min }\left(\sum _{i=1}^n\sum _{j=1}^mG_n^k\left\{i,j\right\}\right)\right]\text{}\mathrm{where}G_n^k=\left(Y_n-S_kY_{n-1}\right)·{\left(Y_n-S_kY_{n-1}\right)}^{*}& \left(46\right)\end{array}$

and Y_n is the received symbol, S_k, k=1, . . . , N are the constellation's matrices and N is the size of the constellation. Applying

Y _n =X _n R+Υ _n

to EQ. 46 we get

G _n ^k=(X_nR+Υ_n−S_k(X_n−1R+σ_n−1))·(X_nR+Υ_n−S_k(X_n−1R+Υ_n−1))*=((S_n−S_k)X_n−1R)·((S_n−S_k)X_n−1)+((S_n−S_k)X_n−1R)·(Υ_n−S_kΥ_n−1)*+((S_n−S_k)X_n−1R)*·(Υ_n−S_kΥ_n−1)+(σ_n−S_kΥ_n−1)·(Υ_n−S_kσ_n−1)*.

9.3. Fast Decoding Algorithm

When trying to decode a symbol sent using two or more antennae in the fashion described above, we use a Maximum Likelihood decoder. In our case, in either known or unknown channels, the computation of the likelihood is quite straightforward. In the known channel case, if the matrix A was transmitted, and the channel coefficients are described by the vector α, then the vector describing the values received, x, will be x=Aα+n, where n is the vector of noise values. Thus, since α is known, the log likelihood of a matrix B will be correlated to

l(B)=∥x−Aα∥²

In the case of an unknown channel, we have A_t+1=AA_i, so x_t+1=AA_tα+n_t+1=A(x_t−n_t) and the log likelihood of a matrix B will be correlated to

l(B)=∥x_t+1−Ax_t∥²

In either case we see that l(B)=∥x−Ay∥² for some known x, y. We simply find the matrix B that minimalizes this quantity and we are done. However, there might be a large number of matrices, e.g. 512, so that going over all of them can be quite time consuming. We suggest a way of doing this much more quickly. Note that such methods are well known, and we simply show that they can be implemented in this case. In order to implement a fast algorithm we will use hashing. We simply divide the space of all possible x, y into subsets, in such a way that for every point in the space we can quickly determine the subset to which it belongs. If our space is Euclidean, one such division might be by “quantization”, rounding off all the coordinates (after some scaling). For each subset we determine in advance what are all the matrices that might minimalize points in the subset. It is easy to see that if the SNR is reasonable, and the subsets are in the probable region, then small subsets will have a small number of possible matrices. Thus, given x, y, we find the subset to which the point belongs, and test only the possible matrices relevant to that subset. Given a reasonable SNR, this means that our decoding works in constant time.

9.4. Two Transmit Diversity

For two transmit diversity there is a more effective algorithm that estimates the codeword matrix's entries independently [3]. We will concentrate on channels with no receive diversity. The generalization for receive diversity is easily obtained according to this section.

9.3.1. Transmission Algorithm

Assume that

$\begin{array}{cc}{X}_{t-1}=\left(\begin{array}{cc}{x}_{t-1}^1& x_{t-1}^2\\ -x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right)& \left(47\right)\end{array}$

was transmitted at time t−1 and that

$\begin{array}{cc}{X}_t=\left(\begin{array}{cc}{x}_t^1& x_t^2\\ -x_t^{2*}& x_t^{1*}\end{array}\right)=S_t·X_{t-1}& \left(48\right)\end{array}$

was transmitted at time t, where

$S_t=\left(\begin{array}{cc}{s}_t^1& s_t^2\\ -s_t^{2*}& s_t^{1*}\end{array}\right).$

Define the channel response at time n as R_t=R=(α₁ α₂)^T where ^T is the transpose operator. We removed the time dependency (t) due to the quasi-static channel assumption. Let the channel noise be Υ_t=(ν_t¹ ν_t²)^T. The received symbols at time t are Y_t=(y_t¹ y_t²)^T=X_tR+Υ_t.

9.3.2. Decoding Algorithm

Every transmitted matrix can be decoded according to the current and previous received symbols by using the differential scheme. First, we calculate:

s _t ¹ =y _t ¹ y _t−1 ¹ *+y _t ² *y _t−1 ²  (49)

by using the following identities:

y _t ¹=(10)·X₁R+ν_t¹=(10)·S_tX_t−1R+ν_t¹

y _t ²=(01)·X_tR+ν_t²=(01)·S_tX_t−1R+ν_t²

y _t−1 ¹=(10)·X_t−1R+ν_t−1¹

y _t−1 ²=(01)·X_t−1R+ν_t−1²  (50)

We rewrite EQ. 49 using the identities in EQ. 50 as:

$\begin{array}{c}{s}_t^1=\left(\left(\begin{array}{cc}{s}_t^1& s_t^2\end{array}\right)X_{t-1}R+{\upsilon }_t^1\right)·{\left(\left(\begin{array}{cc}{x}_{t-1}^1& x_{t-1}^2\end{array}\right)R+{\upsilon }_{t-1}^1\right)}^{*}+\\ \left(\left(\begin{array}{cc}-x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right)R+{\upsilon }_{t-1}^2\right)·{\left(\left(\begin{array}{cc}-s_t^{2*}& s_t^{1*}\end{array}\right)X_{t-1}R+{\upsilon }_t^2\right)}^{*}\\ =\left(\begin{array}{cc}{s}_t^1& s_t^2\end{array}\right)X_{t-1}{\mathrm{RR}}^{*}\left(\begin{array}{c}{x}_{t-1}^{1*}\\ x_{t-1}^{2*}\end{array}\right)+\left(\begin{array}{cc}-x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right){\mathrm{RR}}^{*}X_{t-1}^{*}\left(\begin{array}{c}-s_t^2\\ s_t^1\end{array}\right)+N_t^1\end{array}$

where N_t¹ is a zero-mean noise given by

$N_1=R^{*}\left(\begin{array}{c}{x}_{t-1}^{1*}\\ x_{t-1}^{2*}\end{array}\right){\upsilon }_t^1+\left(\begin{array}{cc}{s}_t^1& s_t^2\end{array}\right)X_{t-1}R{\upsilon }_{t-1}^{1*}+R^{*}X_{t-1}^{*}\left(\begin{array}{c}-s_t^2\\ s_t^1\end{array}\right){\upsilon }_{t-1}^2+\left(\begin{array}{cc}-x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right)R{\upsilon }_t^{2*}+{\upsilon }_t^1{\upsilon }_{t-1}^{1*}+{\upsilon }_{t-1}^2{\upsilon }_t^{2*}.$

On the other hand

$\left(\begin{array}{cc}{s}_t^1& s_t^2\end{array}\right)X_{t-1}{\mathrm{RR}}^{*}\left(\begin{array}{c}{x}_{t-1}^{1*}\\ x_{t-1}^{2*}\end{array}\right)+\left(\begin{array}{cc}-x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right){\mathrm{RR}}^{*}X_{t-1}^{*}\left(\begin{array}{c}-s_t^2\\ s_t^1\end{array}\right)=s_t^1\left[\left(\begin{array}{cc}{x}_{t-1}^1& x_{t-1}^2\end{array}\right){\mathrm{RR}}^{*}\left(\begin{array}{c}{x}_{t-1}^{1*}\\ x_{t-1}^{2*}\end{array}\right)+\left(\begin{array}{cc}-x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right){\mathrm{RR}}^{*}\left(\begin{array}{c}-x_{t-1}^2\\ x_{t-1}^1\end{array}\right)\right]+s_t^2\left[\left(\begin{array}{cc}-x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right){\mathrm{RR}}^{*}\left(\begin{array}{c}{x}_{t-1}^{1*}\\ x_{t-1}^{2*}\end{array}\right)-\left(\begin{array}{cc}-x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right){\mathrm{RR}}^{*}\left(\begin{array}{c}{x}_{t-1}^{1*}\\ x_{t-1}^{2*}\end{array}\right)\right]=s_t^1\left[\left({x_{t-1}^1}^2+{x_{t-1}^2}^2\right)\left({{\alpha }^1}^2+{{\alpha }^2}^2\right)\right].$

According to Eq. 5,

$X_{t-1}=\prod _{i=0}^{t-1}S_t.$

But |x_t−1¹|²+|x_t−1²|²=det(X_t−1) according to EQ. 47. Then

${x_{t-1}^1}^2+{x_{t-1}^2}^2=\prod _{i=0}^{t-1}\text{det}\left(S_i\right)=1.$

Therefore,

s _t ¹ =s _t ¹(|α¹|²+|α²|²)+N_t¹.  (51)

Similarly we compute

s _t ² =y _t ¹ y _t−1 ² *−y _t ² *y _t−1 ¹.  (52)

Using the above identities we write:

$\begin{array}{c}{s}_t^2=\left(\left(\begin{array}{cc}{s}_t^1& s_t^2\end{array}\right)X_{t-1}R+{\upsilon }_t^1\right)·{\left(\left(\begin{array}{cc}-x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right)R+{\upsilon }_{t-1}^2\right)}^{*}-\\ \left(\left(\begin{array}{cc}{x}_{t-1}^1& x_{t-1}^2\end{array}\right)R+{\upsilon }_{t-1}^1\right)·{\left(\left(\begin{array}{cc}-s_t^{2*}& s_t^{1*}\end{array}\right)X_{t-1}R+{\upsilon }_t^2\right)}^{*}\\ =\left(\begin{array}{cc}{s}_t^1& s_t^2\end{array}\right)X_{t-1}{\mathrm{RR}}^{*}\left(\begin{array}{c}-x_{t-1}^2\\ x_{t-1}^1\end{array}\right)+\left(\begin{array}{cc}{x}_{t-1}^1& x_{t-1}^2\end{array}\right){\mathrm{RR}}^{*}X_{t-1}^{*}\left(\begin{array}{c}{s}_t^2\\ -s_t^1\end{array}\right)+N_t^2\end{array}$ $\mathrm{where}$ $N_2=R^{*}\left(\begin{array}{c}-x_{t-1}^2\\ x_{t-1}^1\end{array}\right){\upsilon }_t^1-R^{*}X_{t-1}^{*}\left(\begin{array}{c}-s_t^2\\ s_t^1\end{array}\right){\upsilon }_{t-1}^1+\left(\begin{array}{cc}{s}_t^1& s_t^2\end{array}\right)X_{t-1}R{\upsilon }_{t-1}^{2*}-\left(\begin{array}{cc}{x}_{t-1}^1& x_{t-1}^2\end{array}\right)R{\upsilon }_t^{2*}+{\upsilon }_t^1{\upsilon }_{t-1}^{2*}-{\upsilon }_{t-1}^1{\upsilon }_t^{2*}.$

Rearranging EQ. 52 we get

$\begin{array}{c}{s}_t^2=s_t^1\left[\left(\begin{array}{cc}{x}_{t-1}^1& x_{t-1}^2\end{array}\right){\mathrm{RR}}^{*}\left(\begin{array}{c}-x_{t-1}^2\\ x_{t-1}^1\end{array}\right)-\left(\begin{array}{cc}{x}_{t-1}^1& x_{t-1}^2\end{array}\right){\mathrm{RR}}^{*}\left(\begin{array}{c}-x_{t-1}^2\\ x_{t-1}^1\end{array}\right)\right]+\\ s_t^2\left[\left(\begin{array}{cc}-x_{t-1}^{2*}& x_{t-1}^{1*}\end{array}\right){\mathrm{RR}}^{*}\left(\begin{array}{c}-x_{t-1}^2\\ x_{t-1}^1\end{array}\right)+\left(\begin{array}{cc}{x}_{t-1}^1& x_{t-1}^2\end{array}\right){\mathrm{RR}}^{*}\left(\begin{array}{c}{x}_{t-1}^{1*}\\ x_{t-1}^{2*}\end{array}\right)\right]\\ =s_t^2\left[\left({x_{t-1}^1}^2+{x_{t-1}^2}^2\right)\left({{\alpha }^1}^2+{{\alpha }^2}^2\right)\right]\\ =s_t^2\left({{\alpha }^1}^2+{{\alpha }^2}^2\right).\end{array}$

Finally,

s _t ² =s _t ²(|α¹|²+|α²|²)+N_t².  (53)

The reconstructed matrix S_t is given then by

$S_t=\left({{\alpha }^1}^2+{{\alpha }^2}^2\right)S_t+\left(\begin{array}{cc}{N}_t^1& N_t^2\\ -N_t^{2*}& N_t^{1*}\end{array}\right).$

Since all the quaternion matrices are normalized to determinant 1, we divide the reconstructed matrix by its determinant and choose the closest code matrix as our estimate.

10. SIMULATION RESULTS

We present here the simulation results for the codes described above. The simulations were made on Rayleigh fading channels with Gaussian white noise. Results of Orthogonal Design codes and the G_mr groups were compared to the literature. No additional error correction codes were implemented.

10.1. Known Channel Results

FIGS. 8, 9 and 10 present the BER performance of Quaternion STBC vs. Orthogonal Design when bit allocation is used.

10.2. Unknown Channel Results

FIGS. 11 and 12 present the BER performance of Quaternion STBC vs. Orthogonal Design for unknown channel, when bit allocation is used.

10.3. Results for Codes with No Bit Allocation

We made a few simulations for quaternion group structure 1, 2, 4 and Orthogonal Design 6PSK constellation for known and unknown channels without bit allocation due to the high complexity of allocating bits for rates, which are not integral. The results of these simulations are shown in FIG. 13.

10.4. Exhaustive Search Results

Results for codes that were constructed by exhaustive search are shown in FIGS. 14 and 15. The results are for known channels, though the codes are also applicable for unknown channels due to their unitary design.

10.5. Results for the Gmr Groups and Cosets

FIGS. 16, 17 and 18 represent the results for 2,3-transmit diversity G_m,r groups (n=2, n=3). The performance for 3-transmit diversity is up to 3 db improvement (for rate 2) in comparison to Super Quaternion methodology or Orthogonal Design (both have 2-transmit diversity), while the performance of the 2-transmit diversity code are considerably inferior.

10.6. Results for the Jmr Groups

FIG. 19 shows results for the group J_1,1 and an extended set that includes the group J_1,1 and eight Super Quaternion codewords in order to increase the bit rate allow bit allocation.

10.7. Results for Two Receive Diversity

FIG. 20 demonstrates improved performance when using two receiver antennas for a few examples.

11. CONCLUSIONS

Space Time Codes are a powerful tool for increasing the data rates of channels and the transmission range. The advantage of STBC as was shown in the simulations is obvious over multipath channels. This invention presents new STBC codes with better diversity that achieve improved performance for fading channels. We have found rate bounds on the diversity of orthogonal and unitary codes and presented codes that approach these bounds. The codes presented in this invention were bit allocated for improved performance. These cods were simulated and reached better results than other known STBC codes.

REFERENCES

• [1] I. N. Herstein, “Noncommutative Rings,” The Carus Mathematical Monographs, Number Fifteen.
• [2] Siavash M. Alamouti, “A Simple Transmit Diversity Technique for Wireless Communications,” IEEE Journal on Select Areas in Communications, Vol. 16, No. 8, October 1998.
• [3] Vahid Tarokh. “A Differential Detection Scheme for Transmit Diversity,” IEEE Journal on Select Areas in Communications, Vol. 18, No. 7, July 2000.
• [4] Vahid Tarokh, Nambi Senshadri' A. R. Calderbank. “Space-Time Codes for High Data Rate Wireless Communication Performance Criterion and Code Construction,” IEEE Transactions on Information Theory, Vol. 44, No. 2, July 1998.
• [5] V. Tarokh, H Jafakhani, and A. R. Calderbank, “Space-time block codes from orthogonal designs,” IEEE Trans. Info. Theory, vol. 45, no. 5, pp. 1456-1467, July 1999.
• [6] Bertrand M. Hochwald, Thomas L. Marzetta “Unitary Space-Time modulation for Multiple-Antenna Communications in Rayleigh Flat fading,” IEEE Transactions on Information Theory, March 2000.
• [7] B. Hochwald and W. Sweldens, “Differential unitary space time modulation,” IEEE Trans. Comm., vol. 48, pp. 2041-2052, December 2000.
• [8] A. Shokrollahi, B. Hassibi, B. M. Hochwlad, and W. Sweldens, “Representation theory for high rate multiple antenna code design,” IEEE Trans. Inform. Theory, vol. 47, no. 6, pp. 2335-2367, September 2001.
• [9] B. Hughes, “Differential space-time modulation,” IEEE Trans. Inform. Theory, vol. 46, pp. 2567-2578, November 2000.
• [10] G. Han and J. Rosenthal, “Unitary constellations with large diversity sum and good diversity product,” Department of Mathematics, University of Notre Dame, October 2002.
• [11] M. Shirvani, “The finite inner automorphism groups of division rings,” Math Proc. Camb. Phil. Soc., vol. 118, pp. 207-213, 1995.
• [12] G. J. Foschini and M. J Gans, “On Limits of Wireless Communications in a Fading Environment when Using Multiple Antennas,” Wireless Personal Communications 6: 311-335, 1998.
• [13] S. A. Amitsur, Finite Subgroups of Division Rings, Trans. Amer. Math. Soc. (1955) 361-386.

TABLE 1
Bit allocation of SQ2,7,2
	Binary 0 =	−0.5000 + 0.5000i	0.5000 − 0.5000i
		−0.5000 − 0.5000i	−0.5000 − 0.5000i
	Binary 1 =	−0.7559 + 0.3780i	0.3780 − 0.3780i
		−0.3780 − 0.3780i	−0.7559 − 0.3780i
	Binary 2 =	−0.3780 + 0.7559i	0.3780 − 0.3780i
		−0.3780 − 0.3780i	−0.3780 − 0.7559i
	Binary 3 =	−0.8165 + 0.4082i	0.4082
		−0.4082	−0.8165 − 0.4082i
	Binary 4 =	−0.3780 + 0.3780i	0.3780 − 0.7559i
		−0.3780 − 0.7559i	−0.3780 − 0.3780i
	Binary 5 =	−0.4082	0.4082 − 0.8165i
		−0.4082 − 0.8165i	−0.4082
	Binary 6 =	−0.4472 + 0.8944i	0
		0	−0.4472 − 0.8944i
	Binary 7 =	−0.8944	0.4472
		−0.4472	−0.8944
	Binary 8 =	−0.3780 + 0.3780i	0.7559 − 0.3780i
		−0.7559 − 0.3780i	−0.3780 − 0.3780i
	Binary 9 =	−0.8165	0.4082 − 0.4082i
		−0.4082 − 0.4082i	−0.8165
	Binary 10 =	−0.4082 + 0.4082i	0.8165
		−0.8165	−0.4082 − 0.4082i
	Binary 11 =	−0.5774 + 0.5774i	0.5774
		−0.5774	−0.5774 − 0.5774i
	Binary 12 =	−0.4082	0.8165 − 0.4082i
		−0.8165 − 0.4082i	−0.4082
	Binary 13 =	−0.5774	0.5774 − 0.5774i
		−0.5774 − 0.5774i	−0.5774
	Binary 14 =	−0.4472	0.8944
		−0.8944	−0.4472
	Binary 15 =	−0.4082 + 0.8165i	0.4082
		−0.4082	−0.4082 − 0.8165i
	Binary 16 =	−0.5774 + 0.5774i	0 − 0.5774i
		−0.5774i	−0.5774 − 0.5774i
	Binary 17 =	−0.8165 + 0.4082i	0 − 0.4082i
		0 − 0.4082i	−0.8165 − 0.4082i
	Binary 18 =	−0.4082 + 0.8165i	0 − 0.4082i
		0 − 0.4082i	−0.4082 − 0.8165i
	Binary 19 =	−0.8944 + 0.4472i	0
		0	−0.8944 − 0.4472i
	Binary 20 =	−0.4082 + 0.4082i	0 − 0.8165i
		0 − 0.8165i	−0.4082 − 0.4082i
	Binary 21 =	−0.4472	0 − 0.8944i
		0 − 0.8944i	−0.4472
	Binary 22 =	0 + 0.4082i	−0.4082 − 0.8165i
		0.4082 − 0.8165i	0 − 0.4082i
	Binary 23 =	0.3780 + 0.3780i	−0.3780 − 0.7559i
		0.3780 − 0.7559i	0.3780 − 0.3780i
	Binary 24 =	0.8944	0.4472
		−0.4472	0.8944
	Binary 25 =	−0.8944	0 − 0.4472i
		0 − 0.4472i	−0.8944
	Binary 26 =	0.4082	−0.8165 − 0.4082i
		0.8165 − 0.4082i	0.4082
	Binary 27 =	0.5774	−0.5774 − 0.5774i
		0.5774 − 0.5774i	0.5774
	Binary 28 =	0	−0.4472 − 0.8944i
		0.4472 − 0.8944i	0
	Binary 29 =	0 − 0.4472i	0 − 0.8944i
		0 − 0.8944i	0 + 0.4472i
	Binary 30 =	0.3780 + 0.3780i	−0.7559 − 0.3780i
		0.7559 − 0.3780i	0.3780 − 0.3780i
	Binary 31 =	0.4082	−0.4082 − 0.8165i
		0.4082 − 0.8165i	0.4082
	Binary 32 =	−0.5774 + 0.5774i	−0.5774
		0.5774	−0.5774 − 0.5774i
	Binary 33 =	−0.8165 + 0.4082i	−0.4082
		0.4082	−0.8165 − 0.4082i
	Binary 34 =	−0.4082	0.4082 + 0.8165i
		−0.4082 + 0.8165i	−0.4082
	Binary 35 =	−0.7559 + 0.3780i	0.3780 + 0.3780i
		−0.3780 + 0.3780i	−0.7559 − 0.3780i
	Binary 36 =	−0.4082 + 0.8165i	−0.4082
		0.4082	−0.4082 − 0.8165i
	Binary 37 =	−0.3780 + 0.7559i	−0.3780 + 0.3780i
		0.3780 + 0.3780i	−0.3780 − 0.7559i
	Binary 38 =	−0.5774	0.5774 + 0.5774i
		−0.5774 + 0.5774i	−0.5774
	Binary 39 =	−0.8165	0.4082 + 0.4082i
		−0.4082 + 0.4082i	−0.8165
	Binary 40 =	−0.4082 + 0.4082i	−0.8165
		0.8165	−0.4082 − 0.4082i
	Binary 41 =	−0.3780 + 0.3780i	0.3780 + 0.7559i
		−0.3780 + 0.7559i	−0.3780 − 0.3780i
	Binary 42 =	−0.3780 + 0.3780i	0.7559 + 0.3780i
		−0.7559 + 0.3780i	−0.3780 − 0.3780i
	Binary 43 =	−0.5000 + 0.5000i	0.5000 + 0.5000i
		−0.5000 + 0.5000i	−0.5000 − 0.5000i
	Binary 44 =	0 + 0.8944i	−0.4472
		0.4472	0 − 0.8944i
	Binary 45 =	−0.3780 + 0.3780i	−0.3780 + 0.7559i
		0.3780 + 0.7559i	−0.3780 − 0.3780i
	Binary 46 =	−0.4082	0.8165 + 0.4082i
		−0.8165 + 0.4082i	−0.4082
	Binary 47 =	−0.3780 + 0.7559i	0.3780 + 0.3780i
		−0.3780 + 0.3780i	−0.3780 − 0.7559i
	Binary 48 =	−0.5000 + 0.5000i	−0.5000 − 0.5000i
		0.5000 − 0.5000i	−0.5000 − 0.5000i
	Binary 49 =	−0.7559 + 0.3780i	−0.3780 − 0.3780i
		0.3780 − 0.3780i	−0.7559 − 0.3780i
	Binary 50 =	−0.3780 + 0.7559i	−0.3780 − 0.3780i
		0.3780 − 0.3780i	−0.3780 − 0.7559i
	Binary 51 =	−0.8165 + 0.4082i	0 + 0.4082i
		0 + 0.4082i	−0.8165 − 0.4082i
	Binary 52 =	−0.3780 + 0.3780i	−0.3780 − 0.7559i
		0.3780 − 0.7559i	−0.3780 − 0.3780i
	Binary 53 =	−0.7559 + 0.3780i	−0.3780 + 0.3780i
		0.3780 + 0.3780i	−0.7559 − 0.3780i
	Binary 54 =	0 + 0.8165i	−0.4082 − 0.4082i
		0.4082 − 0.4082i	0 − 0.8165i
	Binary 55 =	−0.8944	0 + 0.4472i
		0 + 0.4472i	−0.8944
	Binary 56 =	−0.3780 + 0.3780i	−0.7559 − 0.3780i
		0.7559 − 0.3780i	−0.3780 − 0.3780i
	Binary 57 =	−0.4082 + 0.4082i	0 + 0.8165i
		0 + 0.8165i	−0.4082 − 0.4082i
	Binary 58 =	0 + 0.4082i	−0.8165 − 0.4082i
		0.8165 − 0.4082i	0 − 0.4082i
	Binary 59 =	−0.5774 + 0.5774i	0 + 0.5774i
		0 + 0.5774i	−0.5774 − 0.5774i
	Binary 60 =	−0.3780 + 0.3780i	−0.7559 + 0.3780i
		0.7559 + 0.3780i	−0.3780 − 0.3780i
	Binary 61 =	−0.5000 + 0.5000i	−0.5000 + 0.5000i
		0.5000 + 0.5000i	−0.5000 − 0.5000i
	Binary 62 =	0 + 0.5774i	−0.5774 − 0.5774i
		0.5774 − 0.5774i	0 − 0.5774i
	Binary 63 =	−0.4082 + 0.8165i	0 + 0.4082i
		0 + 0.4082i	−0.4082 − 0.8165i
	Binary 64 =	0.5000 − 0.5000i	−0.5000 − 0.5000i
		0.5000 − 0.5000i	0.5000 + 0.5000i
	Binary 65 =	0.3780 − 0.3780i	−0.3780 − 0.7559i
		0.3780 − 0.7559i	0.3780 + 0.3780i
	Binary 66 =	0.4082 − 0.8165i	−0.4082
		0.4082	0.4082 + 0.8165i
	Binary 67 =	0 − 0.8944i	−0.4472
		0.4472	0 + 0.8944i
	Binary 68 =	−0.4082 − 0.8165i	0.4082
		−0.4082	−0.4082 + 0.8165i
	Binary 69 =	−0.7559 − 0.3780i	0.3780 − 0.3780i
		−0.3780 − 0.3780i	−0.7559 + 0.3780i
	Binary 70 =	−0.5774 − 0.5774i	0.5774
		−0.5774	−0.5774 + 0.5774i
	Binary 71 =	−0.8165 − 0.4082i	0.4082
		−0.4082	−0.8165 + 0.4082i
	Binary 72 =	0.3780 − 0.7559i	−0.3780 − 0.3780i
		0.3780 − 0.3780i	0.3780 + 0.7559i
	Binary 73 =	0 − 0.8944i	0 − 0.4472i
		0 − 0.4472i	0 + 0.8944i
	Binary 74 =	0.3780 − 0.3780i	−0.7559 − 0.3780i
		0.7559 − 0.3780i	0.3780 + 0.3780i
	Binary 75 =	0 − 0.8165i	−0.4082 − 0.4082i
		0.4082 − 0.4082i	0 + 0.8165i
	Binary 76 =	−0.3780 − 0.3780i	0.7559 − 0.3780i
		−0.7559 − 0.3780i	−0.3780 + 0.3780i
	Binary 77 =	−0.5000 − 0.5000i	0.5000 − 0.5000i
		−0.5000 − 0.5000i	−0.5000 + 0.5000i
	Binary 78 =	−0.4082 − 0.4082i	0.8165
		−0.8165	−0.4082 + 0.4082i
	Binary 79 =	−0.3780 − 0.7559i	0.3780 − 0.3780i
		−0.3780 − 0.3780i	−0.3780 + 0.7559i
	Binary 80 =	−0.5000 − 0.5000i	−0.5000 − 0.5000i
		0.5000 − 0.5000i	−0.5000 + 0.5000i
	Binary 81 =	−0.7559 − 0.3780i	−0.3780 − 0.3780i
		0.3780 − 0.3780i	−0.7559 + 0.3780i
	Binary 82 =	−0.3780 − 0.7559i	−0.3780 − 0.3780i
		0.3780 − 0.3780i	−0.3780 + 0.7559i
	Binary 83 =	−0.4082 − 0.8165i	−0.4082
		0.4082	−0.4082 + 0.8165i
	Binary 84 =	−0.5774 − 0.5774i	0 − 0.5774i
		0 − 0.5774i	−0.5774 + 0.5774i
	Binary 85 =	−0.8165 − 0.4082i	0 − 0.4082i
		0 − 0.4082i	−0.8165 + 0.4082i
	Binary 86 =	−0.4082 − 0.8165i	0 + 0.4082i
		0 + 0.4082i	−0.4082 + 0.8165i
	Binary 87 =	−0.8944 − 0.4472i	0
		0	−0.8944 + 0.4472i
	Binary 88 =	−0.3780 − 0.3780i	−0.3780 − 0.7559i
		0.3780 − 0.7559i	−0.3780 + 0.3780i
	Binary 89 =	−0.3780 − 0.3780i	−0.7559 − 0.3780i
		0.7559 − 0.3780i	−0.3780 + 0.3780i
	Binary 90 =	0 − 0.4082i	−0.8165 − 0.4082i
		0.8165 − 0.4082i	0 + 0.4082i
	Binary 91 =	0 − 0.5774i	−0.5774 − 0.5774i
		0.5774 − 0.5774i	0 + 0.5774i
	Binary 92 =	−0.4082 − 0.4082i	0 − 0.8165i
		0 − 0.8165i	−0.4082 + 0.4082i
	Binary 93 =	−0.3780 − 0.3780i	0.3780 − 0.7559i
		−0.3780 − 0.7559i	−0.3780 + 0.3780i
	Binary 94 =	−0.4082 − 0.8165i	0 − 0.4082i
		0 − 0.4082i	−0.4082 + 0.8165i
	Binary 95 =	−0.4472 − 0.8944i	0
		0	−0.4472 + 0.8944i
	Binary 96 =	0.7559 − 0.3780i	−0.3780 − 0.3780i
		0.3780 − 0.3780i	0.7559 + 0.3780i
	Binary 97 =	−0.8944	−0.4472
		0.4472	−0.8944
	Binary 98 =	−0.3780 − 0.3780i	0.3780 + 0.7559i
		−0.3780 + 0.7559i	−0.3780 + 0.3780i
	Binary 99 =	−0.8165 − 0.4082i	−0.4082
		0.4082	−0.8165 + 0.4082i
	Binary 100 =	−0.3780 − 0.7559i	0.3780 + 0.3780i
		−0.3780 + 0.3780i	−0.3780 + 0.7559i
	Binary 101 =	−0.7559 − 0.3780i	−0.3780 + 0.3780i
		0.3780 + 0.3780i	−0.7559 + 0.3780i
	Binary 102 =	−0.5000 − 0.5000i	0.5000 + 0.5000i
		−0.5000 + 0.5000i	−0.5000 + 0.5000i
	Binary 103 =	−0.7559 − 0.3780i	0.3780 + 0.3780i
		−0.3780 + 0.3780i	−0.7559 + 0.3780i
	Binary 104 =	−0.4472	−0.8944
		0.8944	−0.4472
	Binary 105 =	−0.4472	0 + 0.8944i
		0 + 0.8944i	−0.4472
	Binary 106 =	0.5774 − 0.5774i	−0.5774
		0.5774	0.5774 + 0.5774i
	Binary 107 =	0.8165 − 0.4082i	−0.4082
		0.4082	0.8165 + 0.4082i
	Binary 108 =	−0.3780 − 0.3780i	−0.7559 + 0.3780i
		0.7559 + 0.3780i	−0.3780 + 0.3780i
	Binary 109 =	−0.4082	−0.4082 + 0.8165i
		0.4082 + 0.8165i	−0.4082
	Binary 110 =	−0.3780 − 0.3780i	0.7559 + 0.3780i
		−0.7559 + 0.3780i	−0.3780 + 0.3780i
	Binary 111 =	−0.5000 − 0.5000i	−0.5000 + 0.5000i
		0.5000 + 0.5000i	−0.5000 + 0.5000i
	Binary 112 =	−0.5774	−0.5774 − 0.5774i
		0.5774 − 0.5774i	−0.5774
	Binary 113 =	−0.8165	−0.4082 − 0.4082i
		0.4082 − 0.4082i	−0.8165
	Binary 114 =	−0.4082 − 0.4082i	0 + 0.8165i
		0 + 0.8165i	−0.4082 + 0.4082i
	Binary 115 =	−0.5774 − 0.5774i	−0.5774
		0.5774	−0.5774 + 0.5774i
	Binary 116 =	−0.4082	−0.4082 − 0.8165i
		0.4082 − 0.8165i	−0.4082
	Binary 117 =	−0.8165	−0.4082 + 0.4082i
		0.4082 + 0.4082i	−0.8165
	Binary 118 =	−0.5774 − 0.5774i	0 + 0.5774i
		0 + 0.5774i	−0.5774 + 0.5774i
	Binary 119 =	−0.8165 − 0.4082i	0 + 0.4082i
		0 + 0.4082i	−0.8165 + 0.4082i
	Binary 120 =	−0.4082	−0.8165 − 0.4082i
		0.8165 − 0.4082i	−0.4082
	Binary 121 =	−0.4082 − 0.4082i	−0.8165
		0.8165	−0.4082 + 0.4082i
	Binary 122 =	0	−0.8944 − 0.4472i
		0.8944 − 0.4472i	0
	Binary 123 =	0 − 0.4082i	−0.4082 − 0.8165i
		0.4082 − 0.8165i	0 + 0.4082i
	Binary 124 =	−0.4082	−0.8165 + 0.4082i
		0.8165 + 0.4082i	−0.4082
	Binary 125 =	−0.5774	−0.5774 + 0.5774i
		0.5774 + 0.5774i	−0.5774
	Binary 126 =	−0.3780 − 0.7559i	−0.3780 + 0.3780i
		0.3780 + 0.3780i	−0.3780 + 0.7559i
	Binary 127 =	−0.3780 − 0.3780i	−0.3780 + 0.7559i
		0.3780 + 0.7559i	−0.3780 + 0.3780i
	Binary 128 =	0 + 0.5774i	0.5774 − 0.5774i
		−0.5774 − 0.5774i	0 − 0.5774i
	Binary 129 =	0.7559 + 0.3780i	0.3780 + 0.3780i
		−0.3780 + 0.3780i	0.7559 − 0.3780i
	Binary 130 =	0 + 0.8165i	0.4082 − 0.4082i
		−0.4082 − 0.4082i	0 − 0.8165i
	Binary 126 =	0.3780 + 0.3780i	0.7559 − 0.3780i
		−0.7559 − 0.3780i	0.3780 − 0.3780i
	Binary 126 =	0 + 0.4082i	0.4082 − 0.8165i
		−0.4082 − 0.8165i	0 − 0.4082i
	Binary 127 =	0	0.4472 − 0.8944i
		−0.4472 − 0.8944i	0
	Binary 128 =	0.8165 + 0.4082i	−0.4082
		0.4082	0.8165 − 0.4082i
	Binary 129 =	0.5774 + 0.5774i	0.5774
		−0.5774	0.5774 − 0.5774i
	Binary 130 =	0 + 0.4082i	0.8165 − 0.4082i
		−0.8165 − 0.4082i	0 − 0.4082i
	Binary 131 =	0.5000 + 0.5000i	0.5000 + 0.5000i
		−0.5000 + 0.5000i	0.5000 − 0.5000i
	Binary 132 =	0 + 0.4472i	0.8944
		−0.8944	0 − 0.4472i
	Binary 133 =	0.3780 + 0.3780i	0.7559 + 0.3780i
		−0.7559 + 0.3780i	0.3780 − 0.3780i
	Binary 134 =	0	0.8944 − 0.4472i
		−0.8944 − 0.4472i	0
	Binary 135 =	0.3780 + 0.7559i	0.3780 + 0.3780i
		−0.3780 + 0.3780i	0.3780 − 0.7559i
	Binary 136 =	0.5774 + 0.5774i	−0.5774
		0.5774	0.5774 − 0.5774i
	Binaiy 137 =	0 + 0.8944i	0.4472
		−0.4472	0 − 0.8944i
	Binary 138 =	0.3780 + 0.3780i	0.3780 − 0.7559i
		−0.3780 − 0.7559i	0.3780 − 0.3780i
	Binary 139 =	0.4082 + 0.4082i	0 − 0.8165i
		0 − 0.8165i	0.4082 − 0.4082i
	Binary 140 =	0 + 0.8944i	0 − 0.4472i
		0 − 0.4472i	0 − 0.8944i
	Binary 141 =	0.8165	−0.4082 − 0.4082i
		0.4082 − 0.4082i	0.8165
	Binary 142 =	0 + 0.4472i	0 − 0.8944i
		0 − 0.8944i	0 − 0.4472i
	Binary 143 =	0.4472	0 − 0.8944i
		0 − 0.8944i	0.4472
	Binary 144 =	0.5000 + 0.5000i	−0.5000 − 0.5000i
		0.5000 − 0.5000i	0.5000 − 0.5000i
	Binary 145 =	0.7559 + 0.3780i	−0.3780 − 0.3780i
		0.3780 − 0.3780i	0.7559 − 0.3780i
	Binary 146 =	0.5000 + 0.5000i	−0.5000 + 0.5000i
		0.5000 + 0.5000i	0.5000 − 0.5000i
	Binary 147 =	0.3780 + 0.3780i	−0.3780 + 0.7559i
		0.3780 + 0.7559i	0.3780 − 0.3780i
	Binary 148 =	0.4082	−0.8165 + 0.4082i
		0.8165 + 0.4082i	0.4082
	Binary 149 =	0.5774	−0.5774 + 0.5774i
		0.5774 + 0.5774i	0.5774
	Binary 150 =	0.3780 + 0.3780i	−0.7559 + 0.3780i
		0.7559 + 0.3780i	0.3780 − 0.3780i
	Binary 151 =	0 − 0.4082i	0.4082 − 0.8165i
		−0.4082 − 0.8165i	0 + 0.4082i
	Binary 152 =	0.4082 + 0.4082i	−0.8165
		0.8165	0.4082 − 0.4082i
	Binary 153 =	0.4082 + 0.4082i	0.8165
		−0.8165	0.4082 − 0.4082i
	Binary 154 =	0.3780 + 0.7559i	0.3780 − 0.3780i
		−0.3780 − 0.3780i	0.3780 − 0.7559i
	Binary 155 =	0.5000 + 0.5000i	0.5000 − 0.5000i
		−0.5000 − 0.5000i	0.5000 − 0.5000i
	Binary 156 =	0	0.4472 + 0.8944i
		−0.4472 + 0.8944i	0
	Binary 157 =	0.7559 + 0.3780i	0.3780 − 0.3780i
		−0.3780 − 0.3780i	0.7559 − 0.3780i
	Binary 158 =	0.4472 + 0.8944i	0
		0	0.4472 − 0.8944i
	Binary 159 =	0.4082 + 0.8165i	0.4082
		−0.4082	0.4082 − 0.8165i
	Binary 160 =	0.4082 + 0.8165i	−0.4082
		0.4082	0.4082 − 0.8165i
	Binary 161 =	0.4082 + 0.8165i	0 + 0.4082i
		0 + 0.4082i	0.4082 − 0.8165i
	Binary 162 =	0.7559 + 0.3780i	−0.3780 + 0.3780i
		0.3780 + 0.3780i	0.7559 − 0.3780i
	Binary 163 =	0 + 0.4082i	0.4082 + 0.8165i
		−0.4082 + 0.8165i	0 − 0.4082i
	Binary 164 =	0 + 0.4082i	0.8165 + 0.4082i
		−0.8165 + 0.4082i	0 − 0.4082i
	Binary 165 =	0 + 0.5774i	0.5774 + 0.5774i
		−0.5774 + 0.5774i	0 − 0.5774i
	Binary 166 =	0.3780 + 0.7559i	−0.3780 + 0.3780i
		0.3780 + 0.3780i	0.3780 − 0.7559i
	Binary 167 =	0 + 0.8165i	−0.4082 + 0.4082i
		0.4082 + 0.4082i	0 − 0.8165i
	Binary 168 =	0	0.8944 + 0.4472i
		−0.8944 + 0.4472i	0
	Binary 169 =	0 + 0.8165i	0.4082 + 0.4082i
		−0.4082 + 0.4082i	0 − 0.8165i
	Binary 170 =	0.4082 + 0.8165i	0 − 0.4082i
		0 − 0.4082i	0.4082 − 0.8165i
	Binary 171 =	0.5774 + 0.5774i	0 − 0.5774i
		0 − 0.5774i	0.5774 − 0.5774i
	Binary 172 =	0.8944	−0.4472
		0.4472	0.8944
	Binary 173 =	0.8165 + 0.4082i	0 − 0.4082i
		0 − 0.4082i	0.8165 − 0.4082i
	Binary 174 =	0.5774	0.5774 − 0.5774i
		−0.5774 − 0.5774i	0.5774
	Binary 175 =	0.4082	0.4082 − 0.8165i
		−0.4082 − 0.8165i	0.4082
	Binary 176 =	0.3780 + 0.7559i	−0.3780 − 0.3780i
		0.3780 − 0.3780i	0.3780 − 0.7559i
	Binary 177 =	0.8165	0.4082 − 0.4082i
		−0.4082 − 0.4082i	0.8165
	Binary 178 =	0 + 0.4472i	−0.8944
		0.8944	0 − 0.4472i
	Binary 179 =	0 + 0.4082i	−0.4082 + 0.8165i
		0.4082 + 0.8165i	0 − 0.4082i
	Binary 180 =	0.7559 − 0.3780i	−0.3780 + 0.3780i
		0.3780 + 0.3780i	0.7559 + 0.3780i
	Binary 181 =	0 + 0.4472i	0 + 0.8944i
		0 + 0.8944i	0 − 0.4472i
	Binary 182 =	0 + 0.4082i	−0.8165 + 0.4082i
		0.8165 + 0.4082i	0 − 0.4082i
	Binary 183 =	0 + 0.5774i	−0.5774 + 0.5774i
		0.5774 + 0.5774i	0 − 0.5774i
	Binary 184 =	0.8165	−0.4082 + 0.4082i
		0.4082 + 0.4082i	0.8165
	Binary 185 =	0 + 0.8944i	0 + 0.4472i
		0 + 0.4472i	0 − 0.8944i
	Binary 186 =	0.4472 − 0.8944i	0
		0	0.4472 + 0.8944i
	Binary 187 =	0.4082 − 0.8165i	0 − 0.4082i
		0 − 0.4082i	0.4082 + 0.8165i
	Binary 188 =	0.3780 − 0.7559i	0.3780 + 0.3780i
		−0.3780 + 0.3780i	0.3780 + 0.7559i
	Binary 189 =	0.5000 − 0.5000i	0.5000 + 0.5000i
		−0.5000 + 0.5000i	0.5000 + 0.5000i
	Binary 190 =	0 − 0.8944i	0.4472
		−0.4472	0 + 0.8944i
	Binary 191 =	0.3780 − 0.7559i	0.3780 − 0.3780i
		−0.3780 − 0.3780i	0.3780 + 0.7559i
	Binary 192 =	0 − 0.8165i	0.4082 + 0.4082i
		−0.4082 + 0.4082i	0 + 0.8165i
	Binary 193 =	0.3780 − 0.3780i	0.7559 + 0.3780i
		−0.7559 + 0.3780i	0.3780 + 0.3780i
	Binary 194 =	0.8944	0 + 0.4472i
		0 + 0.4472i	0.8944
	Binary 195 =	0.8165	0.4082 + 0.4082i
		−0.4082 + 0.4082i	0.8165
	Binary 196 =	0.3780 − 0.3780i	0.3780 + 0.7559i
		−0.3780 + 0.7559i	0.3780 + 0.3780i
	Binary 197 =	0.5774	0.5774 + 0.5774i
		−0.5774 + 0.5774i	0.5774
	Binary 198 =	0 − 0.4082i	0.8165 − 0.4082i
		−0.8165 − 0.4082i	0 + 0.4082i
	Binary 199 =	0 − 0.8165i	0.4082 − 0.4082i
		−0.4082 − 0.4082i	0 + 0.8165i
	Binary 200 =	0 − 0.4472i	0.8944
		−0.8944	0 + 0.4472i
	Binary 201 =	0.4082	0.8165 + 0.4082i
		−0.8165 + 0.4082i	0.4082
	Binary 202 =	0.4082 − 0.8165i	0.4082
		−0.4082	0.4082 + 0.8165i
	Binary 203 =	0.5774 − 0.5774i	0 − 0.5774i
		0 − 0.5774i	0.5774 + 0.5774i
	Binary 204 =	0.4082 − 0.8165i	0 + 0.4082i
		0 + 0.4082i	0.4082 + 0.8165i
	Binary 205 =	0.7559 − 0.3780i	0.3780 + 0.3780i
		−0.3780 + 0.3780i	0.7559 + 0.3780i
	Binary 206 =	0.5774 − 0.5774i	0.5774
		−0.5774	0.5774 + 0.5774i
	Binary 207 =	0.4082 − 0.4082i	0 − 0.8165i
		0 − 0.8165i	0.4082 + 0.4082i
	Binary 208 =	0 − 0.8944i	0 + 0.4472i
		0 + 0.4472i	0 + 0.8944i
	Binary 209 =	0.4082 − 0.4082i	0.8165
		−0.8165	0.4082 + 0.4082i
	Binary 210 =	0.8165 − 0.4082i	0 + 0.4082i
		0 + 0.4082i	0.8165 + 0.4082i
	Binary 211 =	0.4082	−0.4082 + 0.8165i
		0.4082 + 0.8165i	0.4082
	Binary 212 =	0.5774 − 0.5774i	0 + 0.5774i
		0 + 0.5774i	0.5774 + 0.5774i
	Binary 213 =	0.4082 − 0.4082i	0 + 0.8165i
		0 + 0.8165i	0.4082 + 0.4082i
	Binary 214 =	0.8165 − 0.4082i	0.4082
		−0.4082	0.8165 + 0.4082i
	Binary 215 =	0 − 0.5774i	0.5774 − 0.5774i
		−0.5774 − 0.5774i	0 + 0.5774i
	Binary 216 =	0.4472	−0.8944
		0.8944	0.4472
	Binary 217 =	4472	0.8944
		−0.8944	0.4472
	Binary 218 =	0.4082	0.8165 − 0.4082i
		−0.8165 − 0.4082i	0.4082
	Binary 219 =	0.8944	0 − 0.4472i
		0 − 0.4472i	0.8944
	Binary 220 =	0 − 0.4082i	0.4082 + 0.8165i
		−0.4082 + 0.8165i	0 + 0.4082i
	Binary 221 =	0.8165 + 0.4082i	0.4082
		−0.4082	0.8165 − 0.4082i
	Binary 222 =	0.3780 − 0.3780i	0.7559 − 0.3780i
		−0.7559 − 0.3780i	0.3780 + 0.3780i
	Binary 223 =	0.5000 − 0.5000i	0.5000 − 0.5000i
		−0.5000 − 0.5000i	0.5000 + 0.5000i
	Binary 224 =	0 − 0.5774i	0.5774 + 0.5774i
		−0.5774 + 0.5774i	0 + 0.5774i
	Binary 225 =	0.8165 + 0.4082i	0 + 0.4082i
		0 + 0.4082i	0.8165 − 0.4082i
	Binary 226 =	0.4082 − 0.4082i	−0.8165
		0.8165	0.4082 + 0.4082i
	Binary 227 =	0.3780 + 0.3780i	0.3780 + 0.7559i
		−0.3780 + 0.7559i	0.3780 − 0.3780i
	Binary 228 =	0.3780 − 0.3780i	−0.7559 + 0.3780i
		0.7559 + 0.3780i	0.3780 + 0.3780i
	Binary 229 =	0.4082	0.4082 + 0.8165i
		−0.4082 + 0.8165i	0.4082
	Binary 230 =	0 − 0.4082i	−0.8165 + 0.4082i
		0.8165 + 0.4082i	0 + 0.4082i
	Binary 231 =	0.4082 + 0.4082i	0 + 0.8165i
		0 + 0.8165i	0.4082 − 0.4082i
	Binary 232 =	0 − 0.4082i	0.8165 + 0.4082i
		−0.8165 + 0.4082i	0 + 0.4082i
	Binary 233 =	0.5774 + 0.5774i	0 + 0.5774i
		0 + 0.5774i	0.5774 − 0.5774i
	Binary 234 =	0.8944 − 0.4472i	0
		0	0.8944 + 0.4472i
	Binary 235 =	0.8165 − 0.4082i	0 − 0.4082i
		0 − 0.4082i	0.8165 + 0.4082i
	Binary 236 =	0.3780 − 0.7559i	−0.3780 + 0.3780i
		0.3780 + 0.3780i	0.3780 + 0.7559i
	Binary 237 =	0.8944 + 0.4472i	0
		0	0.8944 − 0.4472i
	Binary 238 =	0.7559 − 0.3780i	0.3780 − 0.3780i
		−0.3780 − 0.3780i	0.7559 + 0.3780i
	Binary 239 =	0.3780 − 0.3780i	0.3780 − 0.7559i
		−0.3780 − 0.7559i	0.3780 + 0.3780i
	Binary 240 =	0 − 0.8165i	−0.4082 + 0.4082i
		0.4082 + 0.4082i	0 + 0.8165i
	Binary 241 =	0 − 0.4472i	0 + 0.8944i
		0 + 0.8944i	0 + 0.4472i
	Binary 242 =	0 − 0.4472i	−0.8944
		0.8944	0 + 0.4472i
	Binary 243 =	0.4472	0 + 0.8944i
		0 + 0.8944i	0.4472
	Binary 244 =	0.5000 − 0.5000i	−0.5000 + 0.5000i
		0.5000 + 0.5000i	0.5000 + 0.5000i
	Binary 245 =	0.3780 − 0.3780i	−0.3780 + 0.7559i
		0.3780 + 0.7559i	0.3780 + 0.3780i
	Binary 246 =	0	−0.8944 + 0.4472i
		0.8944 + 0.4472i	0
	Binary 247 =	0	−0.4472 + 0.8944i
		0.4472 + 0.8944i	0
	Binary 248 =	0 − 0.5774i	−0.5774 + 0.5774i
		0.5774 + 0.5774i	0 + 0.5774i
	Binary 249 =	0 − 0.4082i	−0.4082 + 0.8165i
		0.4082 + 0.8165i	0 + 0.4082i

TABLE 2
Bit allocation of Coset Extension of G63,37
Binary 0 =	0	1.0000 − 0.0000i	0
	0	0	1.0000 + 0.0000i
	0.7660 + 0.6428i	0	0
Binary 1 =	0.0739 + 0.7609i	−0.0008 + 0.0279i	0.4060 − 0.5000i
	0.1007 − 0.4901i	−0.1545 + 0.6372i	0.4716 − 0.3121i
	−0.4007 + 0.0692i	−0.7469 + 0.1069i	−0.4246 − 0.2916i
Binary 2 =	0.0266 + 0.0083i	0.1062 + 0.6353i	−0.5184 − 0.5619i
	0.5715 + 0.3213i	−0.0781 + 0.5601i	0.2156 + 0.4515I
	−0.0994 + 0.7480i	0.4996 − 0.1254i	0.2776 − 0.2971i
Binary 3 =	0.4351 − 0.2109i	0.5635 − 0.3833i	−0.2743 − 0.4759i
	0.1104 − 0.6464i	−0.0909 − 0.1251i	0.7376 − 0.0445i
	0.2901 + 0.5021i	−0.5223 − 0.4887i	0.2303 − 0.3148i
Binary 4 =	−0.7971 − 0.6038i	0	0
	0	−0.4113 + 0.9115i	0
	0	0	−0.0249 − 0.9997i
Binary 5 =	0.0244 + 0.0134i	−0.5168 + 0.3843i	0.6410 + 0.4166i
	0.4965 + 0.4282i	−0.5341 + 0.1861i	−0.3203 − 0.3843i
	−0.2456 + 0.7135i	0.3395 + 0.3874i	−0.1957 + 0.3564i
Binary 6 =	0.2708 − 0.9626i	0	0
	0	−0.6617 + 0.7498i	0
	0	0	−0.9988 − 0.0498i
Binary 7 =	0.4562 − 0.8899i	0	0
	0	−0.9691 − 0.2468i	0
	0	0	0.9802 − 0.1981i
Binary 8 =	0.2429 + 0.4181i	−0.1510 − 0.6646i	−0.4024 − 0.3739i
	0.6528 + 0.0618i	−0.1497 + 0.0389i	0.6917 − 0.2600i
	−0.4790 + 0.3268i	−0.6457 + 0.3076i	0.1273 − 0.3687i
Binary 9 =	0	0	−0.9010 + 0.4339i
	0.9802 − 0.1981i	0	0
	0	0.4562 − 0.8899i	0
Binary 10 =	−0.2164 − 0.6463i	0.5383 + 0.1093i	−0.2194 + 0.4308i
	−0.1450 + 0.0536i	−0.4547 + 0.5825i	0.3023 + 0.5819i
	−0.6119 + 0.3704i	0.0837 + 0.3810i	−0.5344 − 0.2251i
Binary 11 =	0.2989 − 0.5705i	0.7253 − 0.2414i	−0.0114 − 0.0254i
	0.4004 − 0.3993i	−0.5002 + 0.0108i	−0.1373 − 0.6411i
	−0.4739 − 0.2017i	0.1566 + 0.3752i	0.6266 − 0.4204i
Binary 12 =	−0.7532 + 0.1306i	0.0256 + 0.0109i	0.2406 − 0.5975i
	0.4962 + 0.0638i	0.5367 + 0.3766i	0.3587 − 0.4372i
	−0.0989 − 0.3944i	−0.1734 + 0.7344i	−0.4917 − 0.1535i
Binary 13 =	−0.0840 − 0.6763i	0.1484 + 0.5289i	0.1063 − 0.4717i
	−0.1528 + 0.0238i	−0.7258 − 0.1389i	−0.4365 − 0.4893i
	−0.6731 + 0.2418i	−0.3009 + 0.2483i	0.5734 + 0.0862i
Binary 14 =	0	0.3653 − 0.9309i	0
	0	0	0.9556 − 0.2948i
	−0.5837 + 0.8119i	0	0
Binary 15 =	−0.3631 − 0.4121i	−0.3025 − 0.3772i	−0.3414 + 0.5899i
	0.7142 − 0.1898i	−0.6547 + 0.0362i	0.1362 + 0.0733i
	0.1634 − 0.3542i	0.4249 − 0.3945i	0.6826 + 0.2137i
Binary 16 =	0.0175 − 0.0217i	0.2887 + 0.5757i	0.5216 − 0.5589i
	0.5079 − 0.4146i	0.0905 + 0.5583i	−0.4341 + 0.2487i
	0.6600 + 0.3658i	0.4404 − 0.2671i	0.3170 + 0.2546i
Binary 17 =	0.5425 − 0.8400i	0	0
	0	0.6982 + 0.7159i	0
	0	0	−0.3185 − 0.9479i
Binary 18 =	−0.2594 − 0.7191i	−0.0063 + 0.0272i	0.6302 + 0.1332i
	0.0233 + 0.4998i	−0.2777 + 0.5939i	0.4929 + 0.2773i
	0.3712 − 0.1659i	−0.7533 − 0.0432i	0.0658 − 0.5109i
Binary 19 =	−0.3696 + 0.3118i	0.6283 − 0.2641i	0.2984 − 0.4612i
	0.0525 + 0.6536i	−0.0643 − 0.1406i	0.3761 + 0.6361i
	−0.4050 − 0.4150i	−0.4151 − 0.5825i	0.3852 + 0.0613i
Binary 20 =	−0.6807 − 0.0347i	0.5466 − 0.0543i	−0.4460 − 0.1866i
	−0.0031 + 0.1546i	−0.2628 + 0.6907i	−0.5577 + 0.3449i
	0.1212 + 0.7049i	0.1923 + 0.3394i	0.1845 − 0.5497i
Binary 21 =	0.0230 + 0.0158i	0.6414 − 0.0584i	0.3336 − 0.6878i
	0.4514 + 0.4755i	0.5527 + 0.1197i	−0.3415 + 0.3656i
	−0.3155 + 0.6855i	−0.0877 − 0.5076i	0.3780 + 0.1499i
Binary 22 =	−0.0063 + 0.0272i	0.6302 + 0.1332i	−0.6609 − 0.3841i
	−0.2777 + 0.5939i	0.4929 + 0.2773i	0.3391 + 0.3679i
	−0.7533 − 0.0432i	0.0658 − 0.5109i	0.1777 − 0.3657i
Binary 23 =	−0.0163 − 0.6813i	0.1492 − 0.5287i	−0.4812 − 0.0469i
	−0.1544 + 0.0084i	0.5469 + 0.4970i	−0.4312 + 0.4940i
	−0.6939 + 0.1736i	0.3862 − 0.0550i	0.0143 − 0.5797i
Binary 24 =	−0.1524 + 0.6643i	0.0415 − 0.5477i	−0.2613 + 0.4068i
	0.1517 + 0.0299i	0.6346 + 0.3787i	0.2429 + 0.6091i
	0.7153 + 0.0030i	0.3677 − 0.1304i	−0.5093 − 0.2772i
Binary 25 =	−0.0088 − 0.0265i	−0.2105 + 0.6087i	−0.6959 − 0.3164i
	−0.0728 − 0.6516i	−0.3365 + 0.4546i	0.3740 + 0.3323i
	0.6653 − 0.3559i	0.4987 + 0.1288i	0.1404 − 0.3816i
Binary 26 =	−0.0184 − 0.0210i	0.6171 − 0.1844i	−0.4598 − 0.6107i
	−0.3202 − 0.5722i	0.5655 + 0.0078i	0.1696 + 0.4707i
	0.4748 − 0.5864i	−0.1865 − 0.4801i	0.3058 − 0.2679i
Binary 27 =	0	0	−0.7331 + 0.6802i
	0.2708 + 0.9626i	0	0
	0	−0.6617 − 0.7498i	0
Binary 28 =	−0.0278 − 0.0015i	−0.0218 + 0.6437i	0.2635 − 0.7176i
	−0.6331 − 0.1703i	−0.1875 + 0.5335i	−0.3034 + 0.3978i
	−0.0882 − 0.7494i	0.5145 − 0.0239i	0.3910 + 0.1115i
Binary 29 =	0.3692 + 0.5729i	0.5060 + 0.2138i	−0.4835 + 0.0013i
	0.1273 − 0.0877i	−0.5611 + 0.4809i	−0.3799 + 0.5345i
	0.5016 − 0.5099i	0.0065 + 0.3900i	−0.0435 − 0.5782i
Binary 30 =	0.6283 − 0.2641i	0.2984 − 0.4612i	−0.0827 + 0.4764i
	−0.0643 − 0.1406i	0.3761 + 0.6361i	0.4604 + 0.4670i
	−0.4151 − 0.5825i	0.3852 + 0.0613i	−0.5770 − 0.0576i
Binary 31 =	0.5913 + 0.2554i	0.0403 − 0.7634i	−0.0184 − 0.0210i
	0.4282 + 0.3695i	−0.1727 + 0.4696i	−0.3202 − 0.5722i
	0.1657 − 0.4877i	0.4065 − 0.0087i	0.4748 − 0.5864i
Binary 32 =	−0.5486 + 0.0270i	−0.3005 + 0.3788i	0.0516 − 0.6796i
	0.2969 − 0.6767i	0.1810 + 0.6303i	−0.1545 − 0.0070i
	−0.1751 − 0.3486i	−0.4792 − 0.3265i	−0.7077 + 0.1036i
Binary 33 =	−0.2597 − 0.5894i	−0.3967 − 0.6534i	0.0213 + 0.0180i
	−0.0626 − 0.5621i	0.1219 + 0.4852i	0.4019 + 0.5181i
	−0.4532 + 0.2448i	0.3310 − 0.2362i	−0.3821 + 0.6506i
Binary 34 =	0.3336 − 0.6878i	0.0277 − 0.0027i	0.4538 − 0.4571i
	−0.3415 + 0.3656i	0.6514 + 0.0740i	0.5004 − 0.2636i
	0.3780 + 0.1499i	0.1989 + 0.7279i	−0.3934 − 0.3325i
Binary 35 =	0	0	0.6235 − 0.7818i
	0.4562 − 0.8899i	0	0
	0	−0.9691 − 0.2468i	0
Binary 36 =	0.4828 − 0.0254i	−0.0177 + 0.6813i	−0.4685 − 0.2868i
	0.3528 − 0.5527i	0.1546 − 0.0007i	0.6265 − 0.3919i
	0.0723 + 0.5753i	0.7017 − 0.1388i	0.0517 − 0.3867i
Binary 37 =	0.1191 − 0.6711i	0.5385 − 0.1084i	0.4779 − 0.0733i
	−0.1530 − 0.0223i	−0.1927 + 0.7134i	0.2960 − 0.5851i
	−0.7145 + 0.0326i	0.2251 + 0.3186i	0.1292 + 0.5653i
Binary 38 =	0.4539 − 0.1666i	−0.6807 + 0.0333i	−0.4381 + 0.3313i
	0.1742 − 0.6322i	0.0123 + 0.1541i	−0.1359 − 0.7264i
	0.2386 + 0.5285i	0.1908 + 0.6893i	−0.3410 − 0.1894i
Binary 39 =	0.6138 + 0.1953i	0.3637 + 0.6724i	0.0231 − 0.0156i
	0.4628 + 0.3250i	−0.0975 − 0.4907i	0.6075 − 0.2465i
	0.1164 − 0.5018i	−0.3423 + 0.2194i	0.5228 + 0.5441i
Binary 40 =	−0.4812 − 0.0469i	−0.4504 − 0.5115i	−0.2255 − 0.5009i
	−0.4312 + 0.4940i	−0.1128 + 0.1057i	0.7384 + 0.0291i
	0.0143 − 0.5797i	−0.4200 + 0.5790i	0.2605 − 0.2904i
Binary 41 =	0.5354 − 0.3581i	0.5746 − 0.5042i	−0.0140 + 0.0241i
	0.5427 − 0.1592i	−0.4567 + 0.2043i	−0.4404 + 0.4857i
	−0.3198 − 0.4038i	0.2901 + 0.2849i	−0.7071 − 0.2634i
Binary 42 =	0	−0.1243 − 0.9922i	0
	0	0	0.9950 + 0.0996i
	−0.3185 − 0.9479i	0	0
Binary 43 =	0.4538 − 0.4571i	−0.1865 − 0.7413i	0.0195 − 0.0199i
	0.5004 − 0.2636i	−0.0266 + 0.4996i	0.5466 − 0.3620i
	−0.3934 − 0.3325i	0.3859 − 0.1281i	0.6203 + 0.4297i
Binary 44 =	0.6682 − 0.1343i	0.5465 + 0.0551i	−0.0374 − 0.4821i
	−0.0351 − 0.1506i	−0.3944 + 0.6249i	−0.5614 − 0.3389i
	−0.2914 − 0.6532i	0.1212 + 0.3708i	0.5734 − 0.0866i
Binary 45 =	0.6982 − 0.7159i	0	0
	0	−0.3185 + 0.9479i	0
	0	0	0.5425 + 0.8400i
Binary 46 =	0.0214 − 0.0178i	−0.3806 + 0.5196i	−0.3675 + 0.6703i
	0.5800 − 0.3058i	−0.4555 + 0.3352i	0.3593 − 0.3481i
	0.5744 + 0.4893i	0.4386 + 0.2701i	−0.3700 − 0.1686i
Binary 47 =	0.3428 − 0.4292i	0.3190 − 0.3634i	−0.6402 − 0.2338i
	0.3109 + 0.6704i	−0.1494 − 0.6385i	−0.0485 + 0.1468i
	0.3772 + 0.0994i	0.4623 + 0.3500i	−0.0919 + 0.7093i
Binary 48 =	−0.7365 + 0.2049i	−0.0276 + 0.0041i	0.5629 + 0.3130i
	0.5001 + 0.0141i	−0.6543 − 0.0415i	0.3892 + 0.4103i
	−0.1377 − 0.3826i	−0.2350 − 0.7170i	0.2135 − 0.4688i
Binary 49 =	−0.4113 − 0.9115i	0	0
	0	−0.0249 + 0.9997i	0
	0	0	−0.7971 + 0.6038i
Binary 50 =	−0.4252 − 0.2301i	0.6782 − 0.0671i	−0.4376 − 0.3320i
	−0.5893 + 0.2877i	−0.0200 − 0.1533i	0.6624 − 0.3276i
	0.2383 − 0.5286i	−0.2249 − 0.6790i	0.0899 − 0.3796i
Binary 51 =	0.0257 − 0.0107i	0.4897 + 0.4184i	−0.7582 − 0.0972i
	0.6443 − 0.1213i	0.3002 + 0.4793i	0.4553 + 0.2073i
	0.4046 + 0.6369i	0.3021 − 0.4172i	0.0217 − 0.4060i
Binary 52 =	0.2833 + 0.3918i	0.4754 + 0.4884i	0.4211 − 0.3528i
	0.6557 − 0.0035i	0.1074 − 0.1112i	0.1719 + 0.7187i
	−0.4441 + 0.3729i	0.3906 − 0.5992i	0.3501 + 0.1721i
Binary 53 =	−0.6402 − 0.2338i	−0.0133 − 0.5491i	0.0108 − 0.4834i
	−0.0485 + 0.1468i	0.6691 + 0.3137i	−0.5248 − 0.3931i
	−0.0919 + 0.7093i	0.3529 − 0.1664i	0.5791 − 0.0291i
Binary 54 =	0.9802 + 0.1981i	0	0
	0	0.4562 + 0.8899i	0
	0	0	−0.9691 + 0.2468i
Binary 55 =	0	0	−0.5837 − 0.8119i
	0.3653 + 0.9309i	0	0
	0	0.9556 + 0.2948i	0
Binary 56 =	−0.3984 + 0.5529i	0.4205 + 0.3534i	−0.0349 + 0.4822i
	0.1282 + 0.0865i	−0.6779 + 0.2941i	0.5046 + 0.4188i
	0.6579 + 0.2806i	−0.1087 + 0.3746i	−0.5799 + 0.0002i
Binary 57 =	−0.2255 − 0.5009i	−0.3988 + 0.2734i	−0.6738 − 0.1023i
	0.7384 + 0.0291i	−0.0128 + 0.6556i	−0.0185 + 0.1535i
	0.2605 − 0.2904i	−0.3617 − 0.4532i	0.0504 + 0.7135i
Binary 58 =	0	0	−0.7331 − 0.6802i
	0.9950 + 0.0996i	0	0
	0	−0.8533 − 0.5214i	0
Binary 59 =	−0.5812 − 0.3560i	0.4826 − 0.2623i	0.1088 + 0.4711i
	−0.0767 + 0.1343i	0.0261 + 0.7385i	0.6056 + 0.2515i
	−0.2307 + 0.6770i	0.3090 + 0.2381i	−0.5540 + 0.1711i
Binary 60 =	−0.9397 + 0.3420i	0	0
	0	−0.9397 + 0.3420i	0
	0	0	−0.9397 + 0.3420i
Binary 61 =	0.0194 + 0.0200i	0.3446 + 0.5441i	−0.1536 + 0.7489i
	0.3483 + 0.5555i	0.1456 + 0.5465i	0.2407 − 0.4386i
	−0.4450 + 0.6094i	0.4117 − 0.3096i	−0.4033 − 0.0520i
Binary 62 =	0	0	−0.8533 + 0.5214i
	−0.7331 + 0.6802i	0	0
	0	0.8262 + 0.5633i	0
Binary 63 =	0.3838 − 0.3929i	−0.4616 + 0.1437i	−0.1524 + 0.6643i
	0.2427 + 0.6980i	−0.2055 + 0.6227i	0.1517 + 0.0299i
	0.3655 + 0.1364i	−0.2120 − 0.5397i	0.7153 + 0.0030i
Binary 64 =	0.5989 − 0.3253i	−0.0141 + 0.5491i	−0.4624 − 0.1413i
	−0.0780 − 0.1335i	−0.6526 − 0.3466i	−0.5206 + 0.3987i
	−0.4710 − 0.5383i	−0.3607 + 0.1486i	0.1289 − 0.5654i
Binary 65 =	0.9950 + 0.0996i	0	0
	0	−0.8533 − 0.5214i	0
	0	0	−0.1243 − 0.9922i
Binary 66 =	0.9556 − 0.2948i	0	0
	0	0.0747 + 0.9972i	0
	0	0	0.3653 − 0.9309i
Binary 67 =	0.3553 + 0.3279i	0.6515 − 0.2002i	−0.4951 + 0.2379i
	0.6420 − 0.1334i	−0.0499 − 0.1463i	0.0107 − 0.7389i
	−0.3614 + 0.4535i	−0.3550 − 0.6209i	−0.2968 − 0.2532i
Binary 68 =	−0.0036 + 0.0276i	−0.6072 + 0.2149i	0.4633 − 0.6080i
	−0.2172 + 0.6186i	−0.5652 + 0.0204i	−0.4072 + 0.2907i
	−0.7539 + 0.0320i	0.2102 + 0.4702i	0.3408 + 0.2218i
Binary 69 =	0.6815 + 0.0007i	0.2003 + 0.5115i	0.1552 + 0.4579i
	−0.0046 − 0.1546i	−0.7360 − 0.0659i	0.6276 + 0.1899i
	−0.1562 − 0.6980i	−0.2747 + 0.2770i	−0.5343 + 0.2254i
Binary 70 =	0.7458 − 0.1680i	−0.0036 + 0.0276i	−0.6072 + 0.2149i
	−0.4988 − 0.0390i	−0.2172 + 0.6186i	−0.5652 + 0.0204i
	0.1185 + 0.3890i	−0.7539 + 0.0320i	0.2102 + 0.4702i
Binary 71 =	−0.6617 − 0.7498i	0	0
	0	−0.9988 + 0.0498i	0
	0	0	0.2708 + 0.9626i
Binary 72 =	0	0	−0.9988 + 0.0498i
	−0.4113 + 0.9115i	0	0
	0	−0.0249 − 0.9997i	0
Binary 73 =	0	0	0.5425 − 0.8400i
	0.0747 + 0.9972i	0	0
	0	0.3653 − 0.9309i	0
Binary 74 =	0.5466 − 0.0543i	−0.4460 − 0.1866i	−0.5437 + 0.4109i
	−0.2628 + 0.6907i	−0.5577 + 0.3449i	0.0970 + 0.1204i
	0.1923 + 0.3394i	0.1845 − 0.5497i	0.5460 + 0.4621i
Binary 75 =	0	0	−0.9888 + 0.1490i
	−0.1243 − 0.9922i	0	0
	0	0.9950 + 0.0996i	0
Binary 76 =	0.8262 − 0.5633i	0	0
	0	−0.9888 + 0.1490i	0
	0	0	−0.7331 − 0.6802i
Binary 77 =	0.2497 − 0.6342i	0.3421 + 0.4297i	−0.4450 + 0.1890i
	−0.1456 − 0.0522i	−0.7228 + 0.1540i	−0.1425 + 0.6401i
	−0.7068 − 0.1096i	−0.1808 + 0.3457i	−0.2647 − 0.5159i
Binary 78 =	0.2887 + 0.5757i	0.5216 − 0.5589i	−0.0006 − 0.0279i
	0.0905 + 0.5583i	−0.4341 + 0.2487i	0.1225 − 0.6441i
	0.4404 − 0.2671i	0.3170 + 0.2546i	0.7407 − 0.1440i
Binary 79 =	0.0150 + 0.0235i	−0.3270 + 0.5549i	−0.0359 − 0.7636i
	0.2313 + 0.6135i	−0.4199 + 0.3789i	−0.1251 + 0.4844i
	−0.5569 + 0.5091i	0.4633 + 0.2251i	0.4036 − 0.0491i
Binary 80 =	0.0277 − 0.0027i	0.4538 − 0.4571i	−0.1865 − 0.7413i
	0.6514 + 0.0740i	0.5004 − 0.2636i	−0.0266 + 0.4996i
	0.1989 + 0.7279i	−0.3934 − 0.3325i	0.3859 − 0.1281i
Binary 81 =	0.0614 + 0.4796i	0.6277 + 0.2654i	−0.5322 − 0.1360i
	0.5776 + 0.3105i	0.0558 − 0.1442i	0.4832 − 0.5591i
	−0.5683 + 0.1151i	0.1272 − 0.7039i	−0.0646 − 0.3847i
Binary 82 =	0.0074 + 0.0269i	0.5289 + 0.3675i	0.6977 − 0.3124i
	0.0402 + 0.6544i	0.3465 + 0.4470i	−0.4966 + 0.0606i
	−0.6823 + 0.3223i	0.2591 − 0.4452i	0.1932 + 0.3578i
Binary 83 =	−0.0185 + 0.0208i	0.6324 − 0.1220i	0.0739 + 0.7609i
	−0.5279 + 0.3888i	0.5619 + 0.0641i	0.1007 − 0.4901i
	−0.6409 − 0.3982i	−0.1378 − 0.4963i	−0.4007 + 0.0692i
Binary 84 =	−0.9397 − 0.3420i	0	0
	0	−0.9397 − 0.3420i	0
	0	0	−0.9397 − 0.3420i
Binary 85 =	0	0	−0.6617 + 0.7498i
	−0.7331 − 0.6802i	0	0
	0	0.8262 − 0.5633i	0
Binary 86 =	0.7660 + 0.6428i	0	0
	0	0.7660 + 0.6428i	0
	0	0	0.7660 + 0.6428i
Binary 87 =	0.3971 + 0.5071i	−0.2594 − 0.7191i	−0.0063 + 0.0272i
	0.1993 + 0.5293i	0.0233 + 0.4998i	−0.2777 + 0.5939i
	0.3788 − 0.3490i	0.3712 − 0.1659i	−0.7533 − 0.0432i
Binary 88 =	−0.4504 − 0.5115i	−0.2255 − 0.5009i	−0.3988 + 0.2734i
	−0.1128 + 0.1057i	0.7384 + 0.0291i	−0.0128 + 0.6556i
	−0.4200 + 0.5790i	0.2605 − 0.2904i	−0.3617 − 0.4532i
Binary 89 =	0.3116 − 0.6062i	−0.0679 − 0.5451i	0.2429 + 0.4181i
	−0.1396 − 0.0664i	0.6970 + 0.2455i	0.6528 + 0.0618i
	−0.6924 − 0.1794i	0.3345 − 0.2007i	−0.4790 + 0.3268i
Binary 90 =	0.5627 + 0.3845i	−0.1218 − 0.5356i	−0.4252 − 0.2301i
	0.0833 − 0.1303i	0.7180 + 0.1749i	−0.5893 + 0.2877i
	0.2641 − 0.6647i	0.3129 − 0.2330i	0.2383 − 0.5286i
Binary 91 =	−0.0276 + 0.0041i	0.5629 + 0.3130i	−0.4324 + 0.6304i
	−0.6543 − 0.0415i	0.3892 + 0.4103i	0.3922 − 0.3106i
	−0.2350 − 0.7170i	0.2135 − 0.4688i	−0.3514 − 0.2046i
Binary 92 =	0.0103 − 0.0259i	−0.5525 + 0.3310i	−0.3297 − 0.6897i
	0.3631 − 0.5459i	−0.5499 + 0.1320i	0.0729 + 0.4950i
	0.7385 + 0.1550i	0.2992 + 0.4193i	0.3528 − 0.2020i
Binary 93 =	0.0126 + 0.0249i	0.5684 − 0.3030i	−0.7532 + 0.1306i
	0.1691 + 0.6335i	0.5558 − 0.1044i	0.4962 + 0.0638i
	−0.6049 + 0.4511i	−0.2780 − 0.4337i	−0.0989 − 0.3944i
Binary 94 =	0.0502 + 0.6797i	0.5493 + 0.0004i	0.2833 + 0.3918i
	0.1538 − 0.0161i	−0.3303 + 0.6611i	0.6557 − 0.0035i
	0.6844 − 0.2079i	0.1575 + 0.3569i	−0.4441 + 0.3729i
Binary 95 =	0.4754 + 0.4884i	0.4211 − 0.3528i	0.4689 + 0.1 181i
	0.1074 − 0.1112i	0.1719 + 0.7187i	0.5000 − 0.4242i
	0.3906 − 0.5992i	0.3501 + 0.1721i	−0.1005 + 0.5711i
Binary 96 =	0.0245 − 0.0132i	−0.6360 − 0.1016i	0.1907 − 0.7403i
	0.6291 − 0.1848i	−0.5061 − 0.2524i	−0.2623 + 0.4260i
	0.4661 + 0.5934i	−0.0403 + 0.5135i	0.4002 + 0.0720i
Binary 97 =	−0.6137 − 0.2964i	−0.2750 + 0.4755i	−0.4810 + 0.0494i
	−0.0629 + 0.1412i	−0.4074 − 0.6165i	−0.3248 + 0.5696i
	−0.1621 + 0.6966i	−0.3878 − 0.0420i	−0.1009 − 0.5710i
Binary 98 =	0.6403 + 0.0698i	0.7643 − 0.0169i	−0.0262 + 0.0094i
	0.5180 + 0.2269i	−0.4812 − 0.1371i	−0.6496 + 0.0890i
	0.0146 − 0.5149i	0.0390 + 0.4047i	−0.3724 − 0.6563i
Binary 99 =	0.5746 − 0.5042i	−0.0140 + 0.0241i	0.1799 − 0.6184i
	−0.4567 + 0.2043i	−0.4404 + 0.4857i	0.3134 − 0.4708i
	0.2901 + 0.2849i	−0.7071 − 0.2634i	−0.5045 − 0.1038i
Binary 100 =	−0.0116 + 0.0254i	0.1689 + 0.6215i	0.7353 + 0.2091i
	−0.3899 + 0.5271i	−0.0219 + 0.5651i	−0.4194 − 0.2728i
	−0.7298 − 0.1916i	0.4846 − 0.1745i	−0.0820 + 0.3982i
Binary 101 =	−0.0164 + 0.0226i	−0.4760 + 0.4339i	−0.7642 − 0.0212i
	−0.4866 + 0.4394i	−0.5129 + 0.2383i	0.4737 + 0.1609i
	−0.6774 − 0.3324i	0.3764 + 0.3517i	−0.0188 − 0.4062i
Binary 102 =	0	0	−0.8533 − 0.5214i
	0.5425 − 0.8400i	0	0
	0	0.6982 + 0.7159i	0
Binary 103 =	0.8262 + 0.5633i	0	0
	0	−0.9888 − 0.1490i	0
	0	0	−0.7331 + 0.6802i
Binary 104 =	0.0195 − 0.0199i	0.0538 − 0.6418i	−0.6194 − 0.4480i
	0.5466 − 0.3620i	0.2139 − 0.5235i	0.3008 + 0.3998i
	0.6203 + 0.4297i	−0.5151 − 0.0018i	0.2133 − 0.3462i
Binary 105 =	0	0	−0.6617 − 0.7498i
	−0.7971 − 0.6038i	0	0
	0	−0.4113 + 0.9115i	0
Binary 106 =	0.4255 − 0.5324i	0.4541 − 0.3091i	−0.3385 − 0.3452i
	−0.1237 − 0.0928i	0.0995 + 0.7322i	−0.6479 + 0.1012i
	−0.6431 − 0.3131i	0.3312 + 0.2061i	0.3836 − 0.4349i
Binary 107 =	0.2510 − 0.4886i	−0.1778 − 0.4496i	0.5627 + 0.3845i
	0.4376 + 0.5955i	−0.6363 − 0.1584i	0.0833 − 0.1303i
	0.3894 + 0.0227i	0.5224 − 0.2517i	0.2641 − 0.6647i
Binary 108 =	0.9215 + 0.3884i	0	0
	0	−0.5837 + 0.8119i	0
	0	0	0.8782 − 0.4783i
Binary 109 =	0.5289 + 0.3675i	0.6977 − 0.3124i	0.0230 + 0.0158i
	0.3465 + 0.4470i	−0.4966 + 0.0606i	0.4514 + 0.4755i
	0.2591 − 0.4452i	0.1932 + 0.3578i	−0.3155 + 0.6855i
Binary 110 =	−0.6603 − 0.1689i	0.2977 + 0.4617i	0.4783 + 0.0708i
	−0.0337 + 0.1509i	−0.7345 + 0.0812i	0.4553 − 0.4719i
	−0.0208 + 0.7150i	−0.2143 + 0.3259i	−0.0432 + 0.5782i
Binary 111 =	0.4541 − 0.3091i	−0.3385 − 0.3452i	−0.0163 − 0.6813i
	0.0995 + 0.7322i	−0.6479 + 0.1012i	−0.1544 + 0.0084i
	0.3312 + 0.2061i	0.3836 − 0.4349i	−0.6939 + 0.1736i
Binary 112 =	−0.9988 + 0.0498i	0	0
	0	0.2708 + 0.9626i	0
	0	0	−0.6617 − 0.7498i
Binary 113 =	−0.0262 − 0.0096i	0.6403 + 0.0698i	0.7643 − 0.0169i
	−0.5548 − 0.3494i	0.5180 + 0.2269i	−0.4812 − 0.1371i
	0.1366 − 0.7421i	0.0146 − 0.5149i	0.0390 + 0.4047i
Binary 114 =	−0.0006 − 0.0279i	0.5913 + 0.2554i	0.0403 − 0.7634i
	0.1225 − 0.6441i	0.4282 + 0.3695i	−0.1727 + 0.4696i
	0.7407 − 0.1440i	0.1657 − 0.4877i	0.4065 − 0.0087i
Binary 115 =	0.1839 + 0.6563i	0.2510 − 0.4886i	−0.1778 − 0.4496i
	0.1475 − 0.0463i	0.4376 + 0.5955i	−0.6363 − 0.1584i
	0.6296 − 0.3394i	0.3894 + 0.0227i	0.5224 − 0.2517i
Binary 116 =	−0.2810 + 0.6209i	0.5063 − 0.2130i	0.1528 − 0.4587i
	0.1428 + 0.0594i	−0.0475 + 0.7375i	−0.3856 − 0.5304i
	0.7005 + 0.1447i	0.2838 + 0.2677i	0.5620 + 0.1429i
Binary 117 =	0	0.2708 − 0.9626i	0
	0	0	−0.6617 + 0.7498i
	−0.7331 − 0.6802i	0	0
Binary 118 =	0	0	0.1736 − 0.9848i
	0.7660 − 0.6428i	0	0
	0	0.7660 − 0.6428i	0
Binary 119 =	0.6781 + 0.0686i	0.0958 − 0.5409i	0.4351 − 0.2109i
	0.0108 − 0.1543i	0.5937 + 0.4400i	0.1104 − 0.6464i
	−0.0859 − 0.7101i	0.3788 − 0.0931i	0.2901 + 0.5021i
Binary 120 =	−0.0223 + 0.0167i	0.3971 + 0.5071i	−0.2594 − 0.7191i
	−0.5945 + 0.2765i	0.1993 + 0.5293i	0.0233 + 0.4998i
	−0.5493 − 0.5173i	0.3788 − 0.3490i	0.3712 − 0.1659i
Binary 121 =	0.6782 − 0.0671i	−0.4376 − 0.3320i	−0.4737 + 0.0970i
	−0.0200 − 0.1533i	0.6624 − 0.3276i	−0.2665 + 0.5992i
	−0.2249 − 0.6790i	0.0899 − 0.3796i	−0.1572 − 0.5581i
Binary 122 =	−0.5437 + 0.4109i	0.3838 − 0.3929i	−0.4616 + 0.1437i
	0.0970 + 0.1204i	0.2427 + 0.6980i	−0.2055 + 0.6227i
	0.5460 + 0.4621i	0.3655 + 0.1364i	−0.2120 − 0.5397i
Binary 123 =	−0.0218 + 0.6437i	0.2635 − 0.7176i	−0.0223 + 0.0167i
	−0.1875 + 0.5335i	−0.3034 + 0.3978i	−0.5945 + 0.2765i
	0.5145 − 0.0239i	0.3910 + 0.1115i	−0.5493 − 0.5173i
Binary 124 =	0.6511 + 0.2016i	0.5250 − 0.1615i	−0.3696 + 0.3118i
	0.0412 − 0.1490i	−0.1207 + 0.7291i	0.0525 + 0.6536i
	0.0565 − 0.7130i	0.2557 + 0.2946i	−0.4050 − 0.4150i
Binary 125 =	0	0	−0.7971 − 0.6038i
	−0.9010 + 0.4339i	0	0
	0	0.6235 − 0.7818i	0
Binary 126 =	−0.0237 − 0.0146i	0.2300 + 0.6016i	−0.7365 + 0.2049i
	−0.4746 − 0.4524i	0.0345 + 0.5645i	0.5001 + 0.0141i
	0.2809 − 0.7003i	0.4648 − 0.2219i	−0.1377 − 0.3826i
Binary 127 =	−0.1489 + 0.6266i	0.7588 − 0.0929i	0.0175 − 0.0217i
	−0.2895 + 0.4858i	−0.4924 − 0.0885i	0.5079 − 0.4146i
	0.5091 + 0.0785i	0.0791 + 0.3988i	0.6600 + 0.3658i
Binary 128 =	0.0538 − 0.6418i	−0.6194 − 0.4480i	0.0022 − 0.0278i
	0.2139 − 0.5235i	0.3008 + 0.3998i	0.1860 − 0.6287i
	−0.5151 − 0.0018i	0.2133 − 0.3462i	0.7513 − 0.0695i
Binary 129 =	−0.4947 − 0.2387i	−0.1297 + 0.4658i	−0.5819 + 0.3548i
	0.5844 − 0.4523i	0.4116 + 0.5105i	0.0845 + 0.1295i
	0.0129 − 0.3899i	−0.5684 − 0.1147i	0.4973 + 0.5141i
Binary 130 =	0.4548 + 0.1642i	−0.5428 − 0.4121i	−0.5486 − 0.0278i
	0.5398 − 0.3723i	−0.0897 + 0.1260i	0.3628 − 0.6438i
	−0.1569 + 0.5582i	−0.2969 + 0.6507i	−0.1395 − 0.3643i
Binary 131 =	−0.3005 + 0.3788i	0.0516 − 0.6796i	−0.4030 + 0.3733i
	0.1810 + 0.6303i	−0.1545 − 0.0070i	−0.2076 − 0.7092i
	−0.4792 − 0.3265i	−0.7077 + 0.1036i	−0.3582 − 0.1545i
Binary 132 =	0	0	−0.9397 + 0.3420i
	−0.9397 − 0.3420i	0	0
	0	−0.9397 − 0.3420i	0
Binary 133 =	−0.1322 − 0.4651i	−0.2797 − 0.6215i	−0.5163 + 0.1875i
	−0.6174 − 0.2210i	−0.1390 + 0.0678i	0.0842 − 0.7342i
	0.5448 − 0.1985i	−0.5720 + 0.4295i	−0.2701 − 0.2815i
Binary 134 =	−0.1510 − 0.6646i	−0.4024 − 0.3739i	0.4548 + 0.1642i
	−0.1497 + 0.0389i	0.6917 − 0.2600i	0.5398 − 0.3723i
	−0.6457 + 0.3076i	0.1273 − 0.3687i	−0.1569 + 0.5582i
Binary 135 =	0.0747 − 0.9972i	0	0
	0	0.3653 + 0.9309i	0
	0	0	0.9556 + 0.2948i
Binary 136 =	−0.4616 + 0.1437i	−0.1524 + 0.6643i	0.0415 − 0.5477i
	−0.2055 + 0.6227i	0.1517 + 0.0299i	0.6346 + 0.3787i
	−0.2120 − 0.5397i	0.7153 + 0.0030i	0.3677 − 0.1304i
Binary 137 =	−0.1218 − 0.5356i	−0.4252 − 0.2301i	0.6782 − 0.0671i
	0.7180 + 0.1749i	−0.5893 + 0.2877i	−0.0200 − 0.1533i
	0.3129 − 0.2330i	0.2383 − 0.5286i	−0.2249 − 0.6790i
Binary 138 =	0.0047 + 0.0275i	−0.2597 − 0.5894i	−0.3967 − 0.6534i
	−0.0252 + 0.6552i	−0.0626 − 0.5621i	0.1219 + 0.4852i
	−0.7110 + 0.2528i	−0.4532 + 0.2448i	0.3310 − 0.2362i
Binary 139 =	0.0256 + 0.0109i	0.2406 − 0.5975i	−0.4930 + 0.5842i
	0.5367 + 0.3766i	0.3587 − 0.4372i	0.4212 − 0.2700i
	−0.1734 + 0.7344i	−0.4917 − 0.1535i	−0.3293 − 0.2385i
Binary 140 =	−0.5525 + 0.3310i	−0.3297 − 0.6897i	−0.0088 − 0.0265i
	−0.5499 + 0.1320i	0.0729 + 0.4950i	−0.0728 − 0.6516i
	0.2992 + 0.4193i	0.3528 − 0.2020i	0.6653 − 0.3559i
Binary 141 =	0.6414 − 0.0584i	0.3336 − 0.6878i	0.0277 − 0.0027i
	0.5527 + 0.1197i	−0.3415 + 0.3656i	0.6514 + 0.0740i
	−0.0877 − 0.5076i	0.3780 + 0.1499i	0.1989 + 0.7279i
Binary 142 =	−0.0162 − 0.0227i	−0.6227 − 0.1645i	−0.5488 + 0.5322i
	−0.2616 − 0.6012i	−0.4784 − 0.3015i	0.4460 − 0.2268i
	0.5309 − 0.5362i	−0.0912 + 0.5070i	−0.3039 − 0.2701i
Binary 143 =	0.1492 − 0.5287i	−0.4812 − 0.0469i	−0.4504 − 0.5115i
	0.5469 + 0.4970i	−0.4312 + 0.4940i	−0.1128 + 0.1057i
	0.3862 − 0.0550i	0.0143 − 0.5797i	−0.4200 + 0.5790i
Binary 144 =	0	0	−0.2225 − 0.9749i
	−0.9691 − 0.2468i	0	0
	0	0.9802 − 0.1981i	0
Binary 145 =	0.0128 − 0.0248i	0.6441 + 0.0057i	−0.6433 + 0.4129i
	0.4156 − 0.5071i	0.5381 + 0.1742i	0.4820 − 0.1339i
	0.7194 + 0.2278i	−0.0367 − 0.5138i	−0.2443 − 0.3250i
Binary 146 =	−0.3967 − 0.6534i	0.0213 + 0.0180i	−0.5778 − 0.2846i
	0.1219 + 0.4852i	0.4019 + 0.5181i	−0.4092 − 0.3904i
	0.3310 − 0.2362i	−0.3821 + 0.6506i	−0.1898 + 0.4788i
Binary 147 =	−0.6740 + 0.1009i	0.2011 − 0.5112i	0.3862 + 0.2909i
	0.0276 + 0.1522i	0.4947 + 0.5489i	0.6256 − 0.1967i
	0.2585 + 0.6669i	0.3898 − 0.0162i	−0.3145 + 0.4872i
Binary 148 =	0.7253 − 0.2414i	−0.0114 − 0.0254i	−0.1378 − 0.6292i
	−0.5002 + 0.0108i	−0.1373 − 0.6411i	0.0501 − 0.5633i
	0.1566 + 0.3752i	0.6266 − 0.4204i	−0.4927 + 0.1501i
Binary 149 =	−0.5837 − 0.8119i	0	0
	0	0.8782 + 0.4783i	0
	0	0	0.9215 − 0.3884i
Binary 150 =	0.0277 + 0.0029i	0.6138 + 0.1953i	0.3637 + 0.6724i
	0.6239 + 0.2017i	0.4628 + 0.3250i	−0.0975 − 0.4907i
	0.0508 + 0.7528i	0.1164 − 0.5018i	−0.3423 + 0.2194i
Binary 151 =	0	0.6235 − 0.7818i	0
	0	0	−0.2225 − 0.9749i
	−0.9691 − 0.2468i	0	0
Binary 152 =	−0.6609 − 0.3841i	0.0126 + 0.0249i	0.5684 − 0.3030i
	0.3391 + 0.3679i	0.1691 + 0.6335i	0.5558 − 0.1044i
	0.1777 − 0.3657i	−0.6049 + 0.4511i	−0.2780 − 0.4337i
Binary 153 =	−0.6738 − 0.1023i	−0.4947 − 0.2387i	−0.1297 + 0.4658i
	−0.0185 + 0.1535i	0.5844 − 0.4523i	0.4116 + 0.5105i
	0.0504 + 0.7135i	0.0129 − 0.3899i	−0.5684 − 0.1147i
Binary 154 =	−0.2225 − 0.9749i	0	0
	0	−0.9010 + 0.4339i	0
	0	0	0.6235 − 0.7818i
Binary 155 =	0	0	−0.9691 − 0.2468i
	0.8782 + 0.4783i	0	0
	0	0.9215 − 0.3884i	0
Binary 156 =	−0.6606 + 0.1675i	0.0950 + 0.5410i	−0.3367 + 0.3470i
	0.0426 + 0.1486i	−0.7084 − 0.2104i	0.1173 + 0.6452i
	0.3236 + 0.6379i	−0.3241 + 0.2171i	−0.4443 − 0.3726i
Binary 157 =	−0.0061 − 0.0272i	0.4970 − 0.4096i	−0.2274 + 0.7298i
	−0.0075 − 0.6556i	0.5241 − 0.2124i	0.2832 − 0.4124i
	0.6975 − 0.2879i	−0.3584 − 0.3700i	−0.3961 − 0.0919i
Binary 158 =	0.6277 + 0.2654i	−0.5322 − 0.1360i	0.3553 + 0.3279i
	0.0558 − 0.1442i	0.4832 − 0.5591i	0.6420 − 0.1334i
	0.1272 − 0.7039i	−0.0646 − 0.3847i	−0.3614 + 0.4535i
Binary 159 =	0	0.0747 − 0.9972i	0
	0	0	0.3653 + 0.9309i
	0.5425 + 0.8400i	0	0
Binary 160 =	−0.0251 − 0.0122i	−0.5100 − 0.3934i	−0.1118 − 0.7562i
	−0.5173 − 0.4029i	−0.3237 − 0.4637i	−0.0762 + 0.4945i
	0.2098 − 0.7248i	−0.2810 + 0.4317i	0.3967 − 0.0891i
Binary 161 =	0.2984 − 0.4612i	−0.0827 + 0.4764i	0.3116 − 0.6062i
	0.3761 + 0.6361i	0.4604 + 0.4670i	−0.1396 − 0.0664i
	0.3852 + 0.0613i	−0.5770 − 0.0576i	−0.6924 − 0.1794i
Binary 162 =	−0.4376 − 0.3320i	−0.4737 + 0.0970i	0.4764 − 0.4874i
	0.6624 − 0.3276i	−0.2665 + 0.5992i	−0.1139 − 0.1046i
	0.0899 − 0.3796i	−0.1572 − 0.5581i	−0.6087 − 0.3755i
Binary 163 =	0.0049 − 0.0274i	0.0424 + 0.6427i	0.1493 + 0.7497i
	0.2477 − 0.6070i	−0.1335 + 0.5496i	0.0515 − 0.4976i
	0.7545 + 0.0056i	0.5096 − 0.0750i	−0.3918 + 0.1087i
Binary 164 =	0.7621 + 0.0593i	0.0273 + 0.0056i	−0.4219 − 0.4867i
	−0.4651 − 0.1843i	0.6007 + 0.2628i	−0.2254 − 0.5187i
	−0.0015 + 0.4066i	−0.0244 + 0.7542i	−0.3609 + 0.3675i
Binary 165 =	0	0.1736 − 0.9848i	0
	0	0	0.1736 − 0.9848i
	0.7660 − 0.6428i	0	0
Binary 166 =	−0.6144 + 0.2951i	0.5248 + 0.1623i	0.4119 − 0.2532i
	0.0712 + 0.1373i	−0.5104 + 0.5344i	0.0455 − 0.6542i
	0.4436 + 0.5611i	0.0453 + 0.3875i	0.3386 + 0.4707i
Binary 167 =	0.1799 − 0.6184i	0.1161 − 0.7556i	0.0047 + 0.0275i
	0.3134 − 0.4708i	−0.2186 + 0.4500i	−0.0252 + 0.6552i
	−0.5045 − 0.1038i	0.4054 + 0.0318i	−0.7110 + 0.2528i
Binary 168 =	0	−0.7331 + 0.6802i	0
	0	0	0.8262 + 0.5633i
	−0.6617 − 0.7498i	0	0
Binary 169 =	0.0403 − 0.7634i	−0.0184 − 0.0210i	0.6171 − 0.1844i
	−0.1727 + 0.4696i	−0.3202 − 0.5722i	0.5655 + 0.0078i
	0.4065 − 0.0087i	0.4748 − 0.5864i	−0.1865 − 0.4801i
Binary 170 =	−0.5819 + 0.3548i	−0.5324 + 0.1351i	0.2000 + 0.4402i
	0.0845 + 0.1295i	0.1569 − 0.7221i	0.6434 + 0.1265i
	0.4973 + 0.5141i	−0.2407 − 0.3070i	−0.5092 + 0.2775i
Binary 171 =	0.0516 − 0.6796i	−0.4030 + 0.3733i	0.0133 + 0.4833i
	−0.1545 − 0.0070i	−0.2076 − 0.7092i	0.5438 + 0.3665i
	−0.7077 + 0.1036i	−0.3582 − 0.1545i	−0.5770 + 0.0579i
Binary 172 =	−0.0090 + 0.0264i	−0.4682 − 0.4423i	−0.2989 + 0.7036i
	−0.3354 + 0.5633i	−0.2760 − 0.4936i	0.3228 − 0.3822i
	−0.7453 − 0.1180i	−0.3225 + 0.4016i	−0.3850 − 0.1309i
Binary 173 =	0.6302 + 0.1332i	−0.6609 − 0.3841i	0.0126 + 0.0249i
	0.4929 + 0.2773i	0.3391 + 0.3679i	0.1691 + 0.6335i
	0.0658 − 0.5109i	0.1777 − 0.3657i	−0.6049 + 0.4511i
Binary 174 =	0.0273 − 0.0054i	−0.1489 + 0.6266i	0.7588 − 0.0929i
	0.6556 + 0.0088i	−0.2895 + 0.4858i	−0.4924 − 0.0885i
	0.2704 + 0.7044i	0.5091 + 0.0785i	0.0791 + 0.3988i
Binary 175 =	0	−0.4113 − 0.9115i	0
	0	0	−0.0249 + 0.9997i
	−0.9988 − 0.0498i	0	0
Binary 176 =	0.8782 + 0.4783i	0	0
	0	0.9215 − 0.3884i	0
	0	0	−0.5837 − 0.8119i
Binary 177 =	0.1853 − 0.6559i	−0.5160 − 0.1883i	−0.1322 − 0.4651i
	−0.1500 − 0.0374i	0.5364 − 0.5083i	−0.6174 − 0.2210i
	−0.7142 − 0.0387i	−0.0259 − 0.3892i	0.5448 − 0.1985i
Binary 178 =	0	0	0.2708 − 0.9626i
	−0.9888 + 0.1490i	0	0
	0	−0.7331 − 0.6802i	0
Binary 179 =	0	0	−0.4113 + 0.9115i
	0.6235 − 0.7818i	0	0
	0	−0.2225 − 0.9749i	0
Binary 180 =	0.0020 + 0.0278i	−0.0857 + 0.6384i	−0.5990 + 0.4749i
	−0.0903 + 0.6494i	−0.2397 + 0.5122i	0.4663 − 0.1812i
	−0.7326 + 0.1807i	0.5144 + 0.0275i	−0.2755 − 0.2990i
Binary 181 =	0.0231 − 0.0156i	0.5957 − 0.2449i	0.7108 + 0.2813i
	0.6075 − 0.2465i	0.5635 − 0.0485i	−0.3901 − 0.3132i
	0.5228 + 0.5441i	−0.2334 − 0.4592i	−0.1212 + 0.3881i
Binary 182 =	−0.3401 − 0.5906i	0.3428 − 0.4292i	0.3190 − 0.3634i
	−0.1316 + 0.0813i	0.3109 + 0.6704i	−0.1494 − 0.6385i
	−0.5264 + 0.4843i	0.3772 + 0.0994i	0.4623 + 0.3500i
Binary 183 =	0.3542 − 0.5379i	−0.7448 − 0.1722i	−0.0238 + 0.0144i
	0.4382 − 0.3575i	0.4324 + 0.2516i	−0.6191 + 0.2159i
	−0.4515 − 0.2479i	0.0620 − 0.4018i	−0.4951 − 0.5695i
Binary 184 =	0	−0.9010 − 0.4339i	0
	0	0	0.6235 + 0.7818i
	−0.7971 + 0.6038i	0	0
Binary 185 =	−0.0275 − 0.0043i	0.3542 − 0.5379i	−0.7448 − 0.1722i
	−0.6130 − 0.2325i	0.4382 − 0.3575i	0.4324 + 0.2516i
	−0.0132 − 0.7544i	−0.4515 − 0.2479i	0.0620 − 0.4018i
Binary 186 =	0	0	−0.5837 + 0.8119i
	0.9802 + 0.1981i	0	0
	0	0.4562 + 0.8899i	0
Binary 187 =	−0.5431 − 0.0823i	−0.4742 − 0.0946i	0.1176 + 0.6713i
	0.4251 − 0.6045i	−0.4783 + 0.4486i	0.1514 − 0.0314i
	−0.1026 − 0.3764i	0.0719 − 0.5754i	0.6603 − 0.2750i
Binary 188 =	−0.6407 + 0.2324i	−0.3631 − 0.4121i	−0.3025 − 0.3772i
	0.0572 + 0.1437i	0.7142 − 0.1898i	−0.6547 + 0.0362i
	0.3855 + 0.6025i	0.1634 − 0.3542i	0.4249 − 0.3945i
Binary 189 =	0.3103 + 0.6068i	−0.3203 − 0.4462i	0.0614 + 0.4796i
	0.1354 − 0.0746i	0.7295 − 0.1178i	0.5776 + 0.3105i
	0.5499 − 0.4574i	0.1978 − 0.3362i	−0.5683 + 0.1151i
Binary 190 =	−0.0262 + 0.0094i	0.5354 − 0.3581i	0.5746 − 0.5042i
	−0.6496 + 0.0890i	0.5427 − 0.1592i	−0.4567 + 0.2043i
	−0.3724 − 0.6563i	−0.3198 − 0.4038i	0.2901 + 0.2849i
Binary 191 =	−0.7331 − 0.6802i	0	0
	0	0.8262 − 0.5633i	0
	0	0	−0.9888 + 0.1490i
Binary 192 =	0	0	−0.9988 − 0.0498i
	0.8262 − 0.5633i	0	0
	0	−0.9888 + 0.1490i	0
Binary 193 =	−0.7448 − 0.1722i	−0.0238 + 0.0144i	−0.0744 − 0.6398i
	0.4324 + 0.2516i	−0.6191 + 0.2159i	0.1059 − 0.5555i
	0.0620 − 0.4018i	−0.4951 − 0.5695i	−0.5052 + 0.1003i
Binary 194 =	0.3190 − 0.3634i	−0.6402 − 0.2338i	−0.0133 − 0.5491i
	−0.1494 − 0.6385i	−0.0485 + 0.1468i	0.6691 + 0.3137i
	0.4623 + 0.3500i	−0.0919 + 0.7093i	0.3529 − 0.1664i
Binary 195 =	0.4830 + 0.0228i	0.3704 − 0.5721i	−0.2263 + 0.5005i
	0.4061 − 0.5149i	−0.1323 − 0.0800i	−0.4668 − 0.5729i
	0.0146 + 0.5797i	−0.6711 − 0.2475i	−0.3901 − 0.0032i
Binary 196 =	−0.0855 + 0.6762i	0.2502 + 0.4890i	−0.3712 − 0.3098i
	0.1539 + 0.0147i	−0.7389 + 0.0077i	−0.6346 + 0.1652i
	0.7120 − 0.0682i	−0.2457 + 0.3030i	0.3383 − 0.4709i
Binary 197 =	0.1161 − 0.7556i	0.0047 + 0.0275i	−0.2597 − 0.5894i
	−0.2186 + 0.4500i	−0.0252 + 0.6552i	−0.0626 − 0.5621i
	0.4054 + 0.0318i	−0.7110 + 0.2528i	−0.4532 + 0.2448i
Binary 198 =	−0.2634 − 0.4055i	0.5989 − 0.3253i	−0.0141 + 0.5491i
	−0.6551 − 0.0292i	−0.0780 − 0.1335i	−0.6526 − 0.3466i
	0.4621 − 0.3503i	−0.4710 − 0.5383i	−0.3607 + 0.1486i
Binary 199 =	0	0.9802 − 0.1981i	0
	0	0	0.4562 − 0.8899i
	−0.5837 − 0.8119i	0	0
Binary 200 =	0	0	−0.9010 − 0.4339i
	−0.0249 + 0.9997i	0	0
	0	−0.7971 + 0.6038i	0
Binary 201 =	0	0	−0.1243 − 0.9922i
	0.6982 + 0.7159i	0	0
	0	−0.3185 − 0.9479i	0
Binary 202 =	0.4826 − 0.2623i	0.1088 + 0.4711i	−0.6740 + 0.1009i
	0.0261 + 0.7385i	0.6056 + 0.2515i	0.0276 + 0.1522i
	0.3090 + 0.2381i	−0.5540 + 0.1711i	0.2585 + 0.6669i
Binary 203 =	0.4897 + 0.4184i	−0.7582 − 0.0972i	0.0128 − 0.0248i
	0.3002 + 0.4793i	0.4553 + 0.2073i	0.4156 − 0.5071i
	0.3021 − 0.4172i	0.0217 − 0.4060i	0.7194 + 0.2278i
Binary 204 =	0.0279 + 0.0001i	−0.6256 + 0.1534i	−0.7124 + 0.2772i
	0.6408 + 0.1385i	−0.5644 − 0.0360i	0.4990 − 0.0357i
	0.1255 + 0.7441i	0.1624 + 0.4888i	−0.1751 − 0.3670i
Binary 205 =	0.4456 + 0.4650i	0.5963 + 0.4783i	−0.0262 − 0.0096i
	0.2510 + 0.5068i	−0.2805 − 0.4143i	−0.5548 − 0.3494i
	0.3422 − 0.3850i	−0.2302 + 0.3351i	0.1366 − 0.7421i
Binary 206 =	0	0	−0.9691 + 0.2468i
	0.6235 + 0.7818i	0	0
	0	−0.2225 + 0.9749i	0
Binary 207 =	0.5063 − 0.2130i	0.1528 − 0.4587i	0.1839 + 0.6563i
	−0.0475 + 0.7375i	−0.3856 − 0.5304i	0.1475 − 0.0463i
	0.2838 + 0.2677i	0.5620 + 0.1429i	0.6296 − 0.3394i
Binary 208 =	−0.1243 − 0.9922i	0	0
	0	0.9950 + 0.0996i	0
	0	0	−0.8533 − 0.5214i
Binary 209 =	−0.3025 − 0.3772i	−0.3414 + 0.5899i	−0.5431 − 0.0823i
	−0.6547 + 0.0362i	0.1362 + 0.0733i	0.4251 − 0.6045i
	0.4249 − 0.3945i	0.6826 + 0.2137i	−0.1026 − 0.3764i
Binary 210 =	0.3637 + 0.6724i	0.0231 − 0.0156i	0.5957 − 0.2449i
	−0.0975 − 0.4907i	0.6075 − 0.2465i	0.5635 − 0.0485i
	−0.3423 + 0.2194i	0.5228 + 0.5441i	−0.2334 − 0.4592i
Binary 211 =	0.1977 − 0.4412i	0.1853 − 0.6559i	−0.5160 − 0.1883i
	−0.3309 − 0.5661i	−0.1500 − 0.0374i	0.5364 − 0.5083i
	0.5449 + 0.1982i	−0.7142 − 0.0387i	−0.0259 − 0.3892i
Binary 212 =	−0.4450 + 0.1890i	−0.2164 − 0.6463i	0.5383 + 0.1093i
	−0.1425 + 0.6401i	−0.1450 + 0.0536i	−0.4547 + 0.5825i
	−0.2647 − 0.5159i	−0.6119 + 0.3704i	0.0837 + 0.3810i
Binary 213 =	−0.4737 + 0.0970i	0.4764 − 0.4874i	−0.5486 + 0.0270i
	−0.2665 + 0.5992i	−0.1139 − 0.1046i	0.2969 − 0.6767i
	−0.1572 − 0.5581i	−0.6087 − 0.3755i	−0.1751 − 0.3486i
Binary 214 =	−0.0238 + 0.0144i	−0.0744 − 0.6398i	−0.6813 + 0.3468i
	−0.6191 + 0.2159i	0.1059 − 0.5555i	0.4930 − 0.0853i
	−0.4951 − 0.5695i	−0.5052 + 0.1003i	−0.2108 − 0.3477i
Binary 215 =	0	0	−0.9397 − 0.3420i
	−0.5000 − 0.8660i	0	0
	0	−0.5000 − 0.8660i	0
Binary 216 =	−0.0827 + 0.4764i	0.3116 − 0.6062i	−0.0679 − 0.5451i
	0.4604 + 0.4670i	−0.1396 − 0.0664i	0.6970 + 0.2455i
	−0.5770 − 0.0576i	−0.6924 − 0.1794i	0.3345 − 0.2007i
Binary 217 =	0	0.8782 − 0.4783i	0
	0	0	0.9215 + 0.3884i
	−0.9691 + 0.2468i	0	0
Binary 218 =	0	−0.5837 − 0.8119i	0
	0	0	0.8782 + 0.4783i
	0.9556 + 0.2948i	0	0
Binary 219 =	0.5957 − 0.2449i	0.7108 + 0.2813i	0.0077 − 0.0268i
	0.5635 − 0.0485i	−0.3901 − 0.3132i	0.3069 − 0.5794i
	−0.2334 − 0.4592i	−0.1212 + 0.3881i	0.7502 + 0.0807i
Binary 220 =	0	0.6982 − 0.7159i	0
	0	0	−0.3185 + 0.9479i
	−0.1243 + 0.9922i	0	0
Binary 221 =	0	0	−0.3185 + 0.9479i
	−0.1243 + 0.9922i	0	0
	0	0.9950 − 0.0996i	0
Binary 222 =	0.4764 − 0.4874i	−0.5486 + 0.0270i	−0.3005 + 0.3788i
	−0.1139 − 0.1046i	0.2969 − 0.6767i	0.1810 + 0.6303i
	−0.6087 − 0.3755i	−0.1751 − 0.3486i	−0.4792 − 0.3265i
Binary 223 =	−0.0133 − 0.5491i	0.0108 − 0.4834i	−0.6407 + 0.2324i
	0.6691 + 0.3137i	−0.5248 − 0.3931i	0.0572 + 0.1437i
	0.3529 − 0.1664i	0.5791 − 0.0291i	0.3855 + 0.6025i
Binary 224 =	0.1088 + 0.4711i	−0.6740 + 0.1009i	0.2011 − 0.5112i
	0.6056 + 0.2515i	0.0276 + 0.1522i	0.4947 + 0.5489i
	−0.5540 + 0.1711i	0.2585 + 0.6669i	0.3898 − 0.0162i
Binary 225 =	−0.5000 − 0.8660i	0	0
	0	−0.5000 − 0.8660i	0
	0	0	−0.5000 − 0.8660i
Binary 226 =	0.1176 + 0.6713i	−0.4689 + 0.2861i	−0.4240 + 0.2324i
	0.1514 − 0.0314i	−0.0629 − 0.7363i	−0.0780 + 0.6511i
	0.6603 − 0.2750i	−0.3205 − 0.2224i	−0.3147 − 0.4870i
Binary 227 =	−0.9888 − 0.1490i	0	0
	0	−0.7331 + 0.6802i	0
	0	0	0.8262 + 0.5633i
Binary 228 =	−0.0249 − 0.9997i	0	0
	0	−0.7971 − 0.6038i	0
	0	0	−0.4113 + 0.9115i
Binary 229 =	−0.5160 − 0.1883i	−0.1322 − 0.4651i	−0.2797 − 0.6215i
	0.5364 − 0.5083i	−0.6174 − 0.2210i	−0.1390 + 0.0678i
	−0.0259 − 0.3892i	0.5448 − 0.1985i	−0.5720 + 0.4295i
Binary 230 =	−0.0204 − 0.0190i	−0.4304 + 0.4791i	0.6631 − 0.3803i
	−0.3755 − 0.5374i	−0.4866 + 0.2882i	−0.4881 + 0.1097i
	0.4141 − 0.6308i	0.4095 + 0.3124i	0.2278 + 0.3368i
Binary 231 =	−0.0104 − 0.6440i	0.7621 + 0.0593i	0.0273 + 0.0056i
	0.1607 − 0.5422i	−0.4651 − 0.1843i	0.6007 + 0.2628i
	−0.5127 + 0.0495i	−0.0015 + 0.4066i	−0.0244 + 0.7542i
Binary 232 =	0	0	−0.5000 + 0.8660i
	−0.9397 + 0.3420i	0	0
	0	−0.9397 + 0.3420i	0
Binary 233 =	0.6171 − 0.1844i	−0.4598 − 0.6107i	−0.0275 − 0.0043i
	0.5655 + 0.0078i	0.1696 + 0.4707i	−0.6130 − 0.2325i
	−0.1865 − 0.4801i	0.3058 − 0.2679i	−0.0132 − 0.7544i
Binary 234 =	0.0022 − 0.0278i	−0.3713 − 0.5263i	−0.7625 + 0.0549i
	0.1860 − 0.6287i	−0.1727 − 0.5385i	0.4874 + 0.1129i
	0.7513 − 0.0695i	−0.3957 + 0.3297i	−0.0592 − 0.4023i
Binary 235 =	0.0213 + 0.0180i	−0.5778 − 0.2846i	−0.7239 − 0.2455i
	0.4019 + 0.5181i	−0.4092 − 0.3904i	0.4052 + 0.2934i
	−0.3821 + 0.6506i	−0.1898 + 0.4788i	0.1017 − 0.3937i
Binary 236 =	−0.3973 − 0.5538i	−0.0687 + 0.5450i	0.3209 + 0.3617i
	−0.1228 + 0.0940i	−0.6149 − 0.4099i	0.6521 − 0.0688i
	−0.4755 + 0.5343i	−0.3737 + 0.1119i	−0.4048 + 0.4152i
Binary 237 =	0	0	0.8782 − 0.4783i
	0.4562 + 0.8899i	0	0
	0	−0.9691 + 0.2468i	0
Binary 238 =	−0.4991 − 0.4641i	0.4536 + 0.3098i	−0.2217 − 0.4297i
	−0.1018 + 0.1164i	−0.6453 + 0.3602i	−0.6489 − 0.0942i
	−0.3602 + 0.6179i	−0.0709 + 0.3836i	0.4947 − 0.3025i
Binary 239 =	−0.2750 + 0.4755i	−0.4810 + 0.0494i	−0.6606 + 0.1675i
	−0.4074 − 0.6165i	−0.3248 + 0.5696i	0.0426 + 0.1486i
	−0.3878 − 0.0420i	−0.1009 − 0.5710i	0.3236 + 0.6379i
Binary 240 =	−0.7582 − 0.0972i	0.0128 − 0.0248i	0.6441 + 0.0057i
	0.4553 + 0.2073i	0.4156 − 0.5071i	0.5381 + 0.1742i
	0.0217 − 0.4060i	0.7194 + 0.2278i	−0.0367 − 0.5138i
Binary 241 =	−0.0140 + 0.0241i	0.1799 − 0.6184i	0.1161 − 0.7556i
	−0.4404 + 0.4857i	0.3134 − 0.4708i	−0.2186 + 0.4500i
	−0.7071 − 0.2634i	−0.5045 − 0.1038i	0.4054 + 0.0318i
Binary 242 =	−0.3385 − 0.3452i	−0.0163 − 0.6813i	0.1492 − 0.5287i
	−0.6479 + 0.1012i	−0.1544 + 0.0084i	0.5469 + 0.4970i
	0.3836 − 0.4349i	−0.6939 + 0.1736i	0.3862 − 0.0550i
Binary 243 =	−0.1378 − 0.6292i	0.4005 − 0.6512i	−0.0251 − 0.0122i
	0.0501 − 0.5633i	−0.3762 + 0.3298i	−0.5173 − 0.4029i
	−0.4927 + 0.1501i	0.3612 + 0.1868i	0.2098 − 0.7248i
Binary 244 =	−0.0359 − 0.7636i	0.0266 + 0.0083i	0.1062 + 0.6353i
	−0.1251 + 0.4844i	0.5715 + 0.3213i	−0.0781 + 0.5601i
	0.4036 − 0.0491i	−0.0994 + 0.7480i	0.4996 − 0.1254i
Binary 245 =	−0.5100 − 0.3934i	−0.1118 − 0.7562i	−0.0270 + 0.0068i
	−0.3237 − 0.4637i	−0.0762 + 0.4945i	−0.6552 + 0.0239i
	−0.2810 + 0.4317i	0.3967 − 0.0891i	−0.3052 − 0.6901i
Binary 246 =	−0.4304 + 0.4791i	0.6631 − 0.3803i	−0.0278 − 0.0015i
	−0.4866 + 0.2882i	−0.4881 + 0.1097i	−0.6331 − 0.1703i
	0.4095 + 0.3124i	0.2278 + 0.3368i	−0.0882 − 0.7494i
Binary 247 =	−0.6360 − 0.1016i	0.1907 − 0.7403i	0.0103 − 0.0259i
	−0.5061 − 0.2524i	−0.2623 + 0.4260i	0.3631 − 0.5459i
	−0.0403 + 0.5135i	0.4002 + 0.0720i	0.7385 + 0.1550i
Binary 248 =	−0.3414 + 0.5899i	−0.5431 − 0.0823i	−0.4742 − 0.0946i
	0.1362 + 0.0733i	0.4251 − 0.6045i	−0.4783 + 0.4486i
	0.6826 + 0.2137i	−0.1026 − 0.3764i	0.0719 − 0.5754i
Binary 249 =	0.5635 − 0.3833i	−0.2743 − 0.4759i	0.1977 − 0.4412i
	−0.0909 − 0.1251i	0.7376 − 0.0445i	−0.3309 − 0.5661i
	−0.5223 − 0.4887i	0.2303 − 0.3148i	0.5449 + 0.1982i
Binary 250 =	0.6977 − 0.3124i	0.0230 + 0.0158i	0.6414 − 0.0584i
	−0.4966 + 0.0606i	0.4514 + 0.4755i	0.5527 + 0.1197i
	0.1932 + 0.3578i	−0.3155 + 0.6855i	−0.0877 − 0.5076i
Binary 251 =	0	0.5425 − 0.8400i	0
	0	0	0.6982 + 0.7159i
	0.3653 − 0.9309i	0	0
Binary 252 =	0	0	−0.3185 − 0.9479i
	0.9556 − 0.2948i	0	0
	0	0.0747 + 0.9972i	0
Binary 253 =	0	−0.2225 − 0.9749i	0
	0	0	−0.9010 + 0.4339i
	0.9802 − 0.1981i	0	0
Binary 254 =	−0.3270 + 0.5549i	−0.0359 − 0.7636i	0.0266 + 0.0083i
	−0.4199 + 0.3789i	−0.1251 + 0.4844i	0.5715 + 0.3213i
	0.4633 + 0.2251i	0.4036 − 0.0491i	−0.0994 + 0.7480i
Binary 255 =	0.4536 + 0.3098i	−0.2217 − 0.4297i	−0.6807 − 0.0347i
	−0.6453 + 0.3602i	−0.6489 − 0.0942i	−0.0031 + 0.1546i
	−0.0709 + 0.3836i	0.4947 − 0.3025i	0.1212 + 0.7049i
Binary 256 =	0	−0.9691 + 0.2468i	0
	0	0	0.9802 + 0.1981i
	−0.2225 + 0.9749i	0	0
Binary 257 =	0.4362 + 0.2086i	0.2483 + 0.6347i	0.0406 + 0.5478i
	0.5742 − 0.3167i	0.1422 − 0.0607i	−0.6839 − 0.2799i
	−0.2117 + 0.5398i	0.5927 − 0.4004i	−0.3441 + 0.1837i
Binary 258 =	0.2300 + 0.6016i	−0.7365 + 0.2049i	−0.0276 + 0.0041i
	0.0345 + 0.5645i	0.5001 + 0.0141i	−0.6543 − 0.0415i
	0.4648 − 0.2219i	−0.1377 − 0.3826i	−0.2350 − 0.7170i
Binary 259 =	0	−0.4113 + 0.9115i	0
	0	0	−0.0249 − 0.9997i
	−0.2225 − 0.9749i	0	0
Binary 260 =	0	−0.9988 − 0.0498i	0
	0	0	0.2708 − 0.9626i
	−0.9888 + 0.1490i	0	0
Binary 261 =	−0.1297 + 0.4658i	−0.5819 + 0.3548i	−0.5324 + 0.1351i
	0.4116 + 0.5105i	0.0845 + 0.1295i	0.1569 − 0.7221i
	−0.5684 − 0.1147i	0.4973 + 0.5141i	−0.2407 − 0.3070i
Binary 262 =	0.3862 + 0.2909i	−0.4515 + 0.5105i	−0.1745 − 0.5208i
	0.6256 − 0.1967i	0.1189 + 0.0988i	0.7318 + 0.1025i
	−0.3145 + 0.4872i	0.6267 + 0.3447i	0.2881 − 0.2630i
Binary 263 =	−0.0852 − 0.4759i	−0.0855 + 0.6762i	0.2502 + 0.4890i
	−0.5923 − 0.2813i	0.1539 + 0.0147i	−0.7389 + 0.0077i
	0.5619 − 0.1433i	0.7120 − 0.0682i	−0.2457 + 0.3030i
Binary 264 =	0.0108 − 0.4834i	−0.6407 + 0.2324i	−0.3631 − 0.4121i
	−0.5248 − 0.3931i	0.0572 + 0.1437i	0.7142 − 0.1898i
	0.5791 − 0.0291i	0.3855 + 0.6025i	0.1634 − 0.3542i
Binary 265 =	−0.2177 + 0.6458i	−0.3210 + 0.4457i	0.4362 + 0.2086i
	0.1480 + 0.0449i	−0.3440 − 0.6540i	0.5742 − 0.3167i
	0.7114 + 0.0742i	−0.3817 − 0.0804i	−0.2117 + 0.5398i
Binary 266 =	0.5963 + 0.4783i	−0.0262 − 0.0096i	0.6403 + 0.0698i
	−0.2805 − 0.4143i	−0.5548 − 0.3494i	0.5180 + 0.2269i
	−0.2302 + 0.3351i	0.1366 − 0.7421i	0.0146 − 0.5149i
Binary 267 =	0.1528 − 0.4587i	0.1839 + 0.6563i	0.2510 − 0.4886i
	−0.3856 − 0.5304i	0.1475 − 0.0463i	0.4376 + 0.5955i
	0.5620 + 0.1429i	0.6296 − 0.3394i	0.3894 + 0.0227i
Binary 268 =	0.7353 + 0.2091i	0.0074 + 0.0269i	0.5289 + 0.3675i
	−0.4194 − 0.2728i	0.0402 + 0.6544i	0.3465 + 0.4470i
	−0.0820 + 0.3982i	−0.6823 + 0.3223i	0.2591 − 0.4452i
Binary 269 =	−0.9691 + 0.2468i	0	0
	0	0.9802 + 0.1981i	0
	0	0	0.4562 + 0.8899i
Binary 270 =	−0.0270 + 0.0068i	−0.6435 + 0.0264i	−0.5717 − 0.5075i
	−0.6552 + 0.0239i	−0.5461 − 0.1471i	0.2595 + 0.4278i
	−0.3052 − 0.6901i	0.0623 + 0.5113i	0.2467 − 0.3232i
Binary 271 =	−0.5163 + 0.1875i	−0.4002 − 0.2713i	−0.6137 − 0.2964i
	0.0842 − 0.7342i	−0.6150 + 0.2276i	−0.0629 + 0.1412i
	−0.2701 − 0.2815i	0.2898 − 0.5023i	−0.1621 + 0.6966i
Binary 272 =	0.2003 + 0.5115i	0.1552 + 0.4579i	0.5226 − 0.4375i
	−0.7360 − 0.0659i	0.6276 + 0.1899i	−0.1029 − 0.1154i
	−0.2747 + 0.2770i	−0.5343 + 0.2254i	−0.5683 − 0.4343i
Binary 273 =	−0.2613 + 0.4068i	0.3103 + 0.6068i	−0.3203 − 0.4462i
	0.2429 + 0.6091i	0.1354 − 0.0746i	0.7295 − 0.1178i
	−0.5093 − 0.2772i	0.5499 − 0.4574i	0.1978 − 0.3362i
Binary 274 =	−0.4324 + 0.6304i	−0.0185 + 0.0208i	0.6324 − 0.1220i
	0.3922 − 0.3106i	−0.5279 + 0.3888i	0.5619 + 0.0641i
	−0.3514 − 0.2046i	−0.6409 − 0.3982i	−0.1378 − 0.4963i
Binary 275 =	0.1493 + 0.7497i	−0.0138 − 0.0242i	0.4456 + 0.4650i
	0.0515 − 0.4976i	−0.2004 − 0.6243i	0.2510 + 0.5068i
	−0.3918 + 0.1087i	0.5816 − 0.4807i	0.3422 − 0.3850i
Binary 276 =	−0.1536 + 0.7489i	0.0277 + 0.0029i	0.6138 + 0.1953i
	0.2407 − 0.4386i	0.6239 + 0.2017i	0.4628 + 0.3250i
	−0.4033 − 0.0520i	0.0508 + 0.7528i	0.1164 − 0.5018i
Binary 277 =	−0.7239 − 0.2455i	0.0279 + 0.0001i	−0.6256 + 0.1534i
	0.4052 + 0.2934i	0.6408 + 0.1385i	−0.5644 − 0.0360i
	0.1017 − 0.3937i	0.1255 + 0.7441i	0.1624 + 0.4888i
Binary 278 =	0.0424 + 0.6427i	0.1493 + 0.7497i	−0.0138 − 0.0242i
	−0.1335 + 0.5496i	0.0515 − 0.4976i	−0.2004 − 0.6243i
	0.5096 − 0.0750i	−0.3918 + 0.1087i	0.5816 − 0.4807i
Binary 279 =	0.4119 − 0.2532i	−0.2810 + 0.6209i	0.5063 − 0.2130i
	0.0455 − 0.6542i	0.1428 + 0.0594i	−0.0475 + 0.7375i
	0.3386 + 0.4707i	0.7005 + 0.1447i	0.2838 + 0.2677i
Binary 280 =	0.0950 + 0.5410i	−0.3367 + 0.3470i	−0.3984 + 0.5529i
	−0.7084 − 0.2104i	0.1173 + 0.6452i	0.1282 + 0.0865i
	−0.3241 + 0.2171i	−0.4443 − 0.3726i	0.6579 + 0.2806i
Binary 281 =	0	0	−0.5000 − 0.8660i
	0.1736 − 0.9848i	0	0
	0	0.1736 − 0.9848i	0
Binary 282 =	0.4689 + 0.1181i	0.6781 + 0.0686i	0.0958 − 0.5409i
	0.5000 − 0.4242i	0.0108 − 0.1543i	0.5937 + 0.4400i
	−0.1005 + 0.5711i	−0.0859 − 0.7101i	0.3788 − 0.0931i
Binary 283 =	−0.0783 + 0.7604i	−0.0252 + 0.0120i	−0.2701 + 0.5847i
	0.1959 − 0.4604i	−0.6375 + 0.1532i	−0.3800 + 0.4188i
	−0.4064 − 0.0116i	−0.4359 − 0.6159i	0.4834 + 0.1778i
Binary 284 =	0	0	−0.7971 + 0.6038i
	0.2708 − 0.9626i	0	0
	0	−0.6617 + 0.7498i	0
Binary 285 =	−0.0374 − 0.4821i	0.4255 − 0.5324i	0.4541 − 0.3091i
	−0.5614 − 0.3389i	−0.1237 − 0.0928i	0.0995 + 0.7322i
	0.5734 − 0.0866i	−0.6431 − 0.3131i	0.3312 + 0.2061i
Binary 286 =	0	0.9556 − 0.2948i	0
	0	0	0.0747 + 0.9972i
	0.8782 − 0.4783i	0	0
Binary 287 =	−0.4219 − 0.4867i	0.6220 − 0.4445i	0.0245 − 0.0132i
	−0.2254 − 0.5187i	−0.4748 + 0.1578i	0.6291 − 0.1848i
	−0.3609 + 0.3675i	0.2602 + 0.3124i	0.4661 + 0.5934i
Binary 288 =	0.3209 + 0.3617i	−0.6603 − 0.1689i	0.2977 + 0.4617i
	0.6521 − 0.0688i	−0.0337 + 0.1509i	−0.7345 + 0.0812i
	−0.4048 + 0.4152i	−0.0208 + 0.7150i	−0.2143 + 0.3259i
Binary 289 =	0	−0.7971 + 0.6038i	0
	0	0	−0.4113 − 0.9115i
	−0.6617 + 0.7498i	0	0
Binary 290 =	−0.0744 − 0.6398i	−0.6813 + 0.3468i	−0.0090 + 0.0264i
	0.1059 − 0.5555i	0.4930 − 0.0853i	−0.3354 + 0.5633i
	−0.5052 + 0.1003i	−0.2108 − 0.3477i	−0.7453 − 0.1180i
Binary 291 =	−0.4685 − 0.2868i	0.3536 − 0.3298i	0.4244 + 0.5333i
	0.6265 − 0.3919i	−0.0850 − 0.6502i	0.1180 − 0.1000i
	0.0517 − 0.3867i	0.4252 + 0.3943i	0.4483 − 0.5573i
Binary 292 =	−0.5717 − 0.5075i	−0.0164 + 0.0226i	−0.4760 + 0.4339i
	0.2595 + 0.4278i	−0.4866 + 0.4394i	−0.5129 + 0.2383i
	0.2467 − 0.3232i	−0.6774 − 0.3324i	0.3764 + 0.3517i
Binary 293 =	−0.6194 − 0.4480i	0.0022 − 0.0278i	−0.3713 − 0.5263i
	0.3008 + 0.3998i	0.1860 − 0.6287i	−0.1727 − 0.5385i
	0.2133 − 0.3462i	0.7513 − 0.0695i	−0.3957 + 0.3297i
Binary 294 =	0.0958 − 0.5409i	0.4351 − 0.2109i	0.5635 − 0.3833i
	0.5937 + 0.4400i	0.1104 − 0.6464i	−0.0909 − 0.1251i
	0.3788 − 0.0931i	0.2901 + 0.5021i	−0.5223 − 0.4887i
Binary 295 =	−0.0022 + 0.7644i	0.0152 − 0.0234i	−0.5466 − 0.3407i
	0.1490 − 0.4776i	0.4641 − 0.4632i	−0.3683 − 0.4292i
	−0.4056 + 0.0289i	0.6931 + 0.2983i	−0.2366 + 0.4575i
Binary 296 =	0	−0.9988 + 0.0498i	0
	0	0	0.2708 + 0.9626i
	−0.0249 − 0.9997i	0	0
Binary 297 =	0.4783 + 0.0708i	−0.6144 + 0.2951i	0.5248 + 0.1623i
	0.4553 − 0.4719i	0.0712 + 0.1373i	−0.5104 + 0.5344i
	−0.0432 + 0.5782i	0.4436 + 0.5611i	0.0453 + 0.3875i
Binary 298 =	0	−0.9691 − 0.2468i	0
	0	0	0.9802 − 0.1981i
	0.9215 − 0.3884i	0	0
Binary 299 =	−0.5324 + 0.1351i	0.2000 + 0.4402i	−0.2177 + 0.6458i
	0.1569 − 0.7221i	0.6434 + 0.1265i	0.1480 + 0.0449i
	−0.2407 − 0.3070i	−0.5092 + 0.2775i	0.7114 + 0.0742i
Binary 300 =	0.9556 + 0.2948i	0	0
	0	0.0747 − 0.9972i	0
	0	0	0.3653 + 0.9309i
Binary 301 =	−0.2194 + 0.4308i	−0.5812 − 0.3560i	0.4826 − 0.2623i
	0.3023 + 0.5819i	−0.0767 + 0.1343i	0.0261 + 0.7385i
	−0.5344 − 0.2251i	−0.2307 + 0.6770i	0.3090 + 0.2381i
Binary 302 =	0.7108 + 0.2813i	0.0077 − 0.0268i	0.2989 − 0.5705i
	−0.3901 − 0.3132i	0.3069 − 0.5794i	0.4004 − 0.3993i
	−0.1212 + 0.3881i	0.7502 + 0.0807i	−0.4739 − 0.2017i
Binary 303 =	−0.5184 − 0.5619i	0.0257 − 0.0107i	0.4897 + 0.4184i
	0.2156 + 0.4515i	0.6443 − 0.1213i	0.3002 + 0.4793i
	0.2776 − 0.2971i	0.4046 + 0.6369i	0.3021 − 0.4172i
Binary 304 =	−0.1755 + 0.4505i	0.6815 + 0.0007i	0.2003 + 0.5115i
	0.3587 + 0.5489i	−0.0046 − 0.1546i	−0.7360 − 0.0659i
	−0.5541 − 0.1707i	−0.1562 − 0.6980i	−0.2747 + 0.2770i
Binary 305 =	0.2000 + 0.4402i	−0.2177 + 0.6458i	−0.3210 + 0.4457i
	0.6434 + 0.1265i	0.1480 + 0.0449i	−0.3440 − 0.6540i
	−0.5092 + 0.2775i	0.7114 + 0.0742i	−0.3817 − 0.0804i
Binary 306 =	−0.3210 + 0.4457i	0.4362 + 0.2086i	0.2483 + 0.6347i
	−0.3440 − 0.6540i	0.5742 − 0.3167i	0.1422 − 0.0607i
	−0.3817 − 0.0804i	−0.2117 + 0.5398i	0.5927 − 0.4004i
Binary 307 =	0.4682 − 0.1205i	0.5982 + 0.3266i	0.3832 + 0.3935i
	0.2363 − 0.6117i	0.0699 − 0.1379i	−0.7038 + 0.2252i
	0.1848 + 0.5496i	0.1966 − 0.6877i	−0.1455 + 0.3619i
Binary 308 =	−0.0679 − 0.5451i	0.2429 + 0.4181i	−0.1510 − 0.6646i
	0.6970 + 0.2455i	0.6528 + 0.0618i	−0.1497 + 0.0389i
	0.3345 − 0.2007i	−0.4790 + 0.3268i	−0.6457 + 0.3076i
Binary 309 =	−0.4810 + 0.0494i	−0.6606 + 0.1675i	0.0950 + 0.5410i
	−0.3248 + 0.5696i	0.0426 + 0.1486i	−0.7084 − 0.2104i
	−0.1009 − 0.5710i	0.3236 + 0.6379i	−0.3241 + 0.2171i
Binary 310 =	0	−0.5837 + 0.8119i	0
	0	0	0.8782 − 0.4783i
	0.4562 + 0.8899i	0	0
Binary 311 =	0.1907 − 0.7403i	0.0103 − 0.0259i	−0.5525 + 0.3310i
	−0.2623 + 0.4260i	0.3631 − 0.5459i	−0.5499 + 0.1320i
	0.4002 + 0.0720i	0.7385 + 0.1550i	0.2992 + 0.4193i
Binary 312 =	−0.4460 − 0.1866i	−0.5437 + 0.4109i	0.3838 − 0.3929i
	−0.5577 + 0.3449i	0.0970 + 0.1204i	0.2427 + 0.6980i
	0.1845 − 0.5497i	0.5460 + 0.4621i	0.3655 + 0.1364i
Binary 313 =	−0.3712 − 0.3098i	0.3692 + 0.5729i	0.5060 + 0.2138i
	−0.6346 + 0.1652i	0.1273 − 0.0877i	−0.5611 + 0.4809i
	0.3383 − 0.4709i	0.5016 − 0.5099i	0.0065 + 0.3900i
Binary 314 =	−0.2701 + 0.5847i	0.4288 + 0.6328i	−0.0116 + 0.0254i
	−0.3800 + 0.4188i	−0.1459 − 0.4786i	−0.3899 + 0.5271i
	0.4834 + 0.1778i	−0.3188 + 0.2524i	−0.7298 − 0.1916i
Binary 315 =	−0.6959 − 0.3164i	−0.0237 − 0.0146i	0.2300 + 0.6016i
	0.3740 + 0.3323i	−0.4746 − 0.4524i	0.0345 + 0.5645i
	0.1404 − 0.3816i	0.2809 − 0.7003i	0.4648 − 0.2219i
Binary 316 =	−0.0138 − 0.0242i	0.4456 + 0.4650i	0.5963 + 0.4783i
	−0.2004 − 0.6243i	0.2510 + 0.5068i	−0.2805 − 0.4143i
	0.5816 − 0.4807i	0.3422 − 0.3850i	−0.2302 + 0.3351i
Binary 317 =	0.4633 − 0.6080i	0.0150 + 0.0235i	−0.3270 + 0.5549i
	−0.4072 + 0.2907i	0.2313 + 0.6135i	−0.4199 + 0.3789i
	0.3408 + 0.2218i	−0.5569 + 0.5091i	0.4633 + 0.2251i
Binary 318 =	0.2406 − 0.4194i	−0.5001 + 0.4630i	−0.1226 + 0.5354i
	−0.2729 − 0.5962i	0.1085 + 0.1102i	−0.5710 − 0.4691i
	0.5225 + 0.2514i	0.5893 + 0.4054i	−0.3830 + 0.0741i
Binary 319 =	−0.3203 − 0.4462i	0.0614 + 0.4796i	0.6277 + 0.2654i
	0.7295 − 0.1178i	0.5776 + 0.3105i	0.0558 − 0.1442i
	0.1978 − 0.3362i	−0.5683 + 0.1151i	0.1272 − 0.7039i
Binary 320 =	0.1175 − 0.6333i	0.2949 + 0.7053i	−0.0278 + 0.0013i
	0.2649 − 0.4996i	−0.0482 − 0.4980i	−0.6469 − 0.1064i
	−0.5124 − 0.0531i	−0.3625 + 0.1842i	−0.1624 − 0.7369i
Binary 321 =	−0.7642 − 0.0212i	0.0020 + 0.0278i	−0.0857 + 0.6384i
	0.4737 + 0.1609i	−0.0903 + 0.6494i	−0.2397 + 0.5122i
	−0.0188 − 0.4062i	−0.7326 + 0.1807i	0.5144 + 0.0275i
Binary 322 =	0.5493 + 0.0004i	0.2833 + 0.3918i	0.4754 + 0.4884i
	−0.3303 + 0.6611i	0.6557 − 0.0035i	0.1074 − 0.1112i
	0.1575 + 0.3569i	−0.4441 + 0.3729i	0.3906 − 0.5992i
Binary 323 =	−0.0687 + 0.5450i	0.3209 + 0.3617i	−0.6603 − 0.1689i
	−0.6149 − 0.4099i	0.6521 − 0.0688i	−0.0337 + 0.1509i
	−0.3737 + 0.1119i	−0.4048 + 0.4152i	−0.0208 + 0.7150i
Binary 324 =	−0.6813 + 0.3468i	−0.0090 + 0.0264i	−0.4682 − 0.4423i
	0.4930 − 0.0853i	−0.3354 + 0.5633i	−0.2760 − 0.4936i
	−0.2108 − 0.3477i	−0.7453 − 0.1180i	−0.3225 + 0.4016i
Binary 325 =	0.0133 + 0.4833i	−0.3973 − 0.5538i	−0.0687 + 0.5450i
	0.5438 + 0.3665i	−0.1228 + 0.0940i	−0.6149 − 0.4099i
	−0.5770 + 0.0579i	−0.4755 + 0.5343i	−0.3737 + 0.1119i
Binary 326 =	0.2232 + 0.7311i	0.0244 + 0.0134i	−0.5168 + 0.3843i
	0.0016 − 0.5003i	0.4965 + 0.4282i	−0.5341 + 0.1861i
	−0.3790 + 0.1472i	−0.2456 + 0.7135i	0.3395 + 0.3874i
Binary 327 =	0.2011 − 0.5112i	0.3862 + 0.2909i	−0.4515 + 0.5105i
	0.4947 + 0.5489i	0.6256 − 0.1967i	0.1189 + 0.0988i
	0.3898 − 0.0162i	−0.3145 + 0.4872i	0.6267 + 0.3447i
Binary 328 =	0.0415 − 0.5477i	−0.2613 + 0.4068i	0.3103 + 0.6068i
	0.6346 + 0.3787i	0.2429 + 0.6091i	0.1354 − 0.0746i
	0.3677 − 0.1304i	−0.5093 − 0.2772i	0.5499 − 0.4574i
Binary 329 =	−0.3675 + 0.6703i	0.0049 − 0.0274i	0.0424 + 0.6427i
	0.3593 − 0.3481i	0.2477 − 0.6070i	−0.1335 + 0.5496i
	−0.3700 − 0.1686i	0.7545 + 0.0056i	0.5096 − 0.0750i
Binary 330 =	−0.5828 + 0.2743i	−0.0783 + 0.7604i	−0.0252 + 0.0120i
	−0.5603 + 0.0765i	0.1959 − 0.4604i	−0.6375 + 0.1532i
	0.2560 + 0.4470i	−0.4064 − 0.0116i	−0.4359 − 0.6159i
Binary 331 =	−0.5488 + 0.5322i	−0.0270 − 0.0070i	−0.5828 + 0.2743i
	0.4460 − 0.2268i	−0.5868 − 0.2924i	−0.5603 + 0.0765i
	−0.3039 − 0.2701i	0.0620 − 0.7520i	0.2560 + 0.4470i
Binary 332 =	0	−0.9397 + 0.3420i	0
	0	0	−0.9397 + 0.3420i
	−0.9397 − 0.3420i	0	0
Binary 333 =	0.6410 + 0.4166i	0.0273 − 0.0054i	−0.1489 + 0.6266i
	−0.3203 − 0.3843i	0.6556 + 0.0088i	−0.2895 + 0.4858i
	−0.1957 + 0.3564i	0.2704 + 0.7044i	0.5091 + 0.0785i
Binary 334 =	−0.5466 − 0.3407i	0.4897 + 0.5870i	−0.0034 − 0.0277i
	−0.3683 − 0.4292i	−0.1928 − 0.4617i	0.0578 − 0.6531i
	−0.2366 + 0.4575i	−0.2921 + 0.2829i	0.7227 − 0.2170i
Binary 335 =	−0.4381 + 0.3313i	0.2406 − 0.4194i	−0.5001 + 0.4630i
	−0.1359 − 0.7264i	−0.2729 − 0.5962i	0.1085 + 0.1102i
	−0.3410 − 0.1894i	0.5225 + 0.2514i	0.5893 + 0.4054i
Binary 336 =	−0.1753 + 0.5206i	−0.1755 + 0.4505i	0.6815 + 0.0007i
	−0.5215 − 0.5236i	0.3587 + 0.5489i	−0.0046 − 0.1546i
	−0.3885 + 0.0356i	−0.5541 − 0.1707i	−0.1562 − 0.6980i
Binary 337 =	0.1552 + 0.4579i	0.5226 − 0.4375i	0.4822 + 0.2631i
	0.6276 + 0.1899i	−0.1029 − 0.1154i	−0.6062 + 0.4226i
	−0.5343 + 0.2254i	−0.5683 − 0.4343i	−0.0324 + 0.3888i
Binary 338 =	0.4288 + 0.6328i	−0.0116 + 0.0254i	0.1689 + 0.6215i
	−0.1459 − 0.4786i	−0.3899 + 0.5271i	−0.0219 + 0.5651i
	−0.3188 + 0.2524i	−0.7298 − 0.1916i	0.4846 − 0.1745i
Binary 339 =	0.0406 + 0.5478i	0.4682 − 0.1205i	0.5982 + 0.3266i
	−0.6839 − 0.2799i	0.2363 − 0.6117i	0.0699 − 0.1379i
	−0.3441 + 0.1837i	0.1848 + 0.5496i	0.1966 − 0.6877i
Binary 340 =	−0.3367 + 0.3470i	−0.3984 + 0.5529i	0.4205 + 0.3534i
	0.1173 + 0.6452i	0.1282 + 0.0865i	−0.6779 + 0.2941i
	−0.4443 − 0.3726i	0.6579 + 0.2806i	−0.1087 + 0.3746i
Binary 341 =	0	−0.6617 + 0.7498i	0
	0	0	−0.9988 − 0.0498i
	0.8262 − 0.5633i	0	0
Binary 342 =	0.5250 − 0.1615i	−0.3696 + 0.3118i	0.6283 − 0.2641i
	−0.1207 + 0.7291i	0.0525 + 0.6536i	−0.0643 − 0.1406i
	0.2557 + 0.2946i	−0.4050 − 0.4150i	−0.4151 − 0.5825i
Binary 343 =	0.7524 + 0.1349i	−0.0204 − 0.0190i	−0.4304 + 0.4791i
	−0.4444 − 0.2297i	−0.3755 − 0.5374i	−0.4866 + 0.2882i
	−0.0419 + 0.4044i	0.4141 − 0.6308i	0.4095 + 0.3124i
Binary 344 =	0.4205 + 0.3534i	−0.0349 + 0.4822i	0.0502 + 0.6797i
	−0.6779 + 0.2941i	0.5046 + 0.4188i	0.1538 − 0.0161i
	−0.1087 + 0.3746i	−0.5799 + 0.0002i	0.6844 − 0.2079i
Binary 345 =	1.0000 − 0.0000i	0	0
	0	1.0000 + 0.0000i	0
	0	0	1.0000 − 0.0000i
Binary 346 =	0.2977 + 0.4617i	0.4783 + 0.0708i	−0.6144 + 0.2951i
	−0.7345 + 0.0812i	0.4553 − 0.4719i	0.0712 + 0.1373i
	−0.2143 + 0.3259i	−0.0432 + 0.5782i	0.4436 + 0.5611i
Binary 347 =	0.4897 + 0.5870i	−0.0034 − 0.0277i	−0.6377 + 0.0903i
	−0.1928 − 0.4617i	0.0578 − 0.6531i	−0.5580 − 0.0920i
	−0.2921 + 0.2829i	0.7227 − 0.2170i	0.1129 + 0.5026i
Binary 348 =	−0.3297 − 0.6897i	−0.0088 − 0.0265i	−0.2105 + 0.6087i
	0.0729 + 0.4950i	−0.0728 − 0.6516i	−0.3365 + 0.4546i
	0.3528 − 0.2020i	0.6653 − 0.3559i	0.4987 + 0.1288i
Binary 349 =	0.8782 − 0.4783i	0	0
	0	0.9215 + 0.3884i	0
	0	0	−0.5837 + 0.8119i
Binary 350 =	0.2502 + 0.4890i	−0.3712 − 0.3098i	0.3692 + 0.5729i
	−0.7389 + 0.0077i	−0.6346 + 0.1652i	0.1273 − 0.0877i
	−0.2457 + 0.3030i	0.3383 − 0.4709i	0.5016 − 0.5099i
Binary 351 =	0.5248 + 0.1623i	0.4119 − 0.2532i	−0.2810 + 0.6209i
	−0.5104 + 0.5344i	0.0455 − 0.6542i	0.1428 + 0.0594i
	0.0453 + 0.3875i	0.3386 + 0.4707i	0.7005 + 0.1447i
Binary 352 =	−0.5322 − 0.1360i	0.3553 + 0.3279i	0.6515 − 0.2002i
	0.4832 − 0.5591i	0.6420 − 0.1334i	−0.0499 − 0.1463i
	−0.0646 − 0.3847i	−0.3614 + 0.4535i	−0.3550 − 0.6209i
Binary 353 =	0.5629 + 0.3130i	−0.4324 + 0.6304i	−0.0185 + 0.0208i
	0.3892 + 0.4103i	0.3922 − 0.3106i	−0.5279 + 0.3888i
	0.2135 − 0.4688i	−0.3514 − 0.2046i	−0.6409 − 0.3982i
Binary 354 =	0.4822 + 0.2631i	0.4132 + 0.2510i	0.1191 − 0.6711i
	−0.6062 + 0.4226i	0.6029 − 0.2580i	−0.1530 − 0.0223i
	−0.0324 + 0.3888i	−0.2644 + 0.5161i	−0.7145 + 0.0326i
Binary 355 =	0	−0.9888 + 0.1490i	0
	0	0	−0.7331 − 0.6802i
	0.9950 + 0.0996i	0	0
Binary 356 =	0.6324 − 0.1220i	0.0739 + 0.7609i	−0.0008 + 0.0279i
	0.5619 + 0.0641i	0.1007 − 0.4901i	−0.1545 + 0.6372i
	−0.1378 − 0.4963i	−0.4007 + 0.0692i	−0.7469 + 0.1069i
Binary 357 =	0	0	0.4562 − 0.8899i
	−0.5837 − 0.8119i	0	0
	0	0.8782 + 0.4783i	0
Binary 358 =	0	0	0.0747 − 0.9972i
	−0.3185 + 0.9479i	0	0
	0	0.5425 + 0.8400i	0
Binary 359 =	−0.2263 + 0.5005i	0.3846 − 0.2930i	−0.0840 − 0.6763i
	−0.4668 − 0.5729i	−0.0199 − 0.6554i	−0.1528 + 0.0238i
	−0.3901 − 0.0032i	0.3838 + 0.4346i	−0.6731 + 0.2418i
Binary 360 =	−0.0857 + 0.6384i	−0.5990 + 0.4749i	0.0194 + 0.0200i
	−0.2397 + 0.5122i	0.4663 − 0.1812i	0.3483 + 0.5555i
	0.5144 + 0.0275i	−0.2755 − 0.2990i	−0.4450 + 0.6094i
Binary 361 =	0.4211 − 0.3528i	0.4689 + 0.1181i	0.6781 + 0.0686i
	0.1719 + 0.7187i	0.5000 − 0.4242i	0.0108 − 0.1543i
	0.3501 + 0.1721i	−0.1005 + 0.5711i	−0.0859 − 0.7101i
Binary 362 =	−0.3806 + 0.5196i	−0.3675 + 0.6703i	0.0049 − 0.0274i
	−0.4555 + 0.3352i	0.3593 − 0.3481i	0.2477 − 0.6070i
	0.4386 + 0.2701i	−0.3700 − 0.1686i	0.7545 + 0.0056i
Binary 363 =	−0.0349 + 0.4822i	0.0502 + 0.6797i	0.5493 + 0.0004i
	0.5046 + 0.4188i	0.1538 − 0.0161i	−0.3303 + 0.6611i
	−0.5799 + 0.0002i	0.6844 − 0.2079i	0.1575 + 0.3569i
Binary 364 =	0	0	0.6982 − 0.7159i
	−0.8533 + 0.5214i	0	0
	0	−0.1243 + 0.9922i	0
Binary 365 =	0.2483 + 0.6347i	0.0406 + 0.5478i	0.4682 − 0.1205i
	0.1422 − 0.0607i	−0.6839 − 0.2799i	0.2363 − 0.6117i
	0.5927 − 0.4004i	−0.3441 + 0.1837i	0.1848 + 0.5496i
Binary 366 =	0.6441 + 0.0057i	−0.6433 + 0.4129i	−0.0061 − 0.0272i
	0.5381 + 0.1742i	0.4820 − 0.1339i	−0.0075 − 0.6556i
	−0.0367 − 0.5138i	−0.2443 − 0.3250i	0.6975 − 0.2879i
Binary 367 =	−0.1997 − 0.6123i	−0.0022 + 0.7644i	0.0152 − 0.0234i
	−0.0063 − 0.5655i	0.1490 − 0.4776i	0.4641 − 0.4632i
	−0.4753 + 0.1984i	−0.4056 + 0.0289i	0.6931 + 0.2983i
Binary 368 =	0	−0.9888 − 0.1490i	0
	0	0	−0.7331 + 0.6802i
	0.2708 + 0.9626i	0	0
Binary 369 =	−0.7625 + 0.0549i	−0.0162 − 0.0227i	−0.6227 − 0.1645i
	0.4874 + 0.1129i	−0.2616 − 0.6012i	−0.4784 − 0.3015i
	−0.0592 − 0.4023i	0.5309 − 0.5362i	−0.0912 + 0.5070i
Binary 370 =	−0.6227 − 0.1645i	−0.5488 + 0.5322i	−0.0270 − 0.0070i
	−0.4784 − 0.3015i	0.4460 − 0.2268i	−0.5868 − 0.2924i
	−0.0912 + 0.5070i	−0.3039 − 0.2701i	0.0620 − 0.7520i
Binary 371 =	0	0.9950 − 0.0996i	0
	0	0	−0.8533 + 0.5214i
	−0.7331 + 0.6802i	0	0
Binary 372 =	0	−0.9397 − 0.3420i	0
	0	0	−0.9397 − 0.3420i
	−0.5000 − 0.8660i	0	0
Binary 373 =	−0.6430 − 0.0378i	0.2232 + 0.7311i	0.0244 + 0.0134i
	−0.5287 − 0.2008i	0.0016 − 0.5003i	0.4965 + 0.4282i
	0.0111 + 0.5150i	−0.3790 + 0.1472i	−0.2456 + 0.7135i
Binary 374 =	0	0	1.0000 − 0.0000i
	0.7660 + 0.6428i	0	0
	0	0.7660 + 0.6428i	0
Binary 375 =	0	0	0.7660 − 0.6428i
	1.0000 + 0.0000i	0	0
	0	1.0000 − 0.0000i	0
Binary 376 =	−0.7124 + 0.2772i	0.0214 − 0.0178i	−0.3806 + 0.5196i
	0.4990 − 0.0357i	0.5800 − 0.3058i	−0.4555 + 0.3352i
	−0.1751 − 0.3670i	0.5744 + 0.4893i	0.4386 + 0.2701i
Binary 377 =	0	−0.5000 + 0.8660i	0
	0	0	−0.5000 + 0.8660i
	−0.9397 + 0.3420i	0	0
Binary 378 =	0.3832 + 0.3935i	0.2812 − 0.3933i	0.6682 − 0.1343i
	−0.7038 + 0.2252i	−0.2122 − 0.6205i	−0.0351 − 0.1506i
	−0.1455 + 0.3619i	0.4949 + 0.3022i	−0.2914 − 0.6532i
Binary 379 =	0.5216 − 0.5589i	−0.0006 − 0.0279i	0.5913 + 0.2554i
	−0.4341 + 0.2487i	0.1225 − 0.6441i	0.4282 + 0.3695i
	0.3170 + 0.2546i	0.7407 − 0.1440i	0.1657 − 0.4877i
Binary 380 =	−0.8533 + 0.5214i	0	0
	0	−0.1243 + 0.9922i	0
	0	0	0.9950 − 0.0996i
Binary 381 =	−0.0270 − 0.0070i	−0.5828 + 0.2743i	−0.0783 + 0.7604i
	−0.5868 − 0.2924i	−0.5603 + 0.0765i	0.1959 − 0.4604i
	0.0620 − 0.7520i	0.2560 + 0.4470i	−0.4064 − 0.0116i
Binary 382 =	0.3421 + 0.4297i	−0.4450 + 0.1890i	−0.2164 − 0.6463i
	−0.7228 + 0.1540i	−0.1425 + 0.6401i	−0.1450 + 0.0536i
	−0.1808 + 0.3457i	−0.2647 − 0.5159i	−0.6119 + 0.3704i
Binary 383 =	0	0.8262 − 0.5633i	0
	0	0	−0.9888 + 0.1490i
	−0.1243 − 0.9922i	0	0
Binary 384 =	−0.0278 + 0.0013i	−0.3171 − 0.5606i	0.6793 + 0.3507i
	−0.6469 − 0.1064i	−0.1182 − 0.5530i	−0.3570 − 0.3505i
	−0.1624 − 0.7369i	−0.4266 + 0.2887i	−0.1593 + 0.3741i
Binary 385 =	−0.4689 + 0.2861i	−0.4240 + 0.2324i	0.5216 + 0.4386i
	−0.0629 − 0.7363i	−0.0780 + 0.6511i	0.0958 − 0.1214i
	−0.3205 − 0.2224i	−0.3147 − 0.4870i	0.3290 − 0.6351i
Binary 386 =	0.9802 − 0.1981i	0	0
	0	0.4562 − 0.8899i	0
	0	0	−0.9691 − 0.2468i
Binary 387 =	0	−0.8533 − 0.5214i	0
	0	0	−0.1243 − 0.9922i
	0.6982 + 0.7159i	0	0
Binary 388 =	0.2949 + 0.7053i	−0.0278 + 0.0013i	−0.3171 − 0.5606i
	−0.0482 − 0.4980i	−0.6469 − 0.1064i	−0.1182 − 0.5530i
	−0.3625 + 0.1842i	−0.1624 − 0.7369i	−0.4266 + 0.2887i
Binary 389 =	−0.4742 − 0.0946i	0.1176 + 0.6713i	−0.4689 + 0.2861i
	−0.4783 + 0.4486i	0.1514 − 0.0314i	−0.0629 − 0.7363i
	0.0719 − 0.5754i	0.6603 − 0.2750i	−0.3205 − 0.2224i
Binary 390 =	−0.0252 + 0.0120i	−0.2701 + 0.5847i	0.4288 + 0.6328i
	−0.6375 + 0.1532i	−0.3800 + 0.4188i	−0.1459 − 0.4786i
	−0.4359 − 0.6159i	0.4834 + 0.1778i	−0.3188 + 0.2524i
Binary 391 =	0.4005 − 0.6512i	−0.0251 − 0.0122i	−0.5100 − 0.3934i
	−0.3762 + 0.3298i	−0.5173 − 0.4029i	−0.3237 − 0.4637i
	0.3612 + 0.1868i	0.2098 − 0.7248i	−0.2810 + 0.4317i
Binary 392 =	0.2406 − 0.5975i	−0.4930 + 0.5842i	0.0267 − 0.0081i
	0.3587 − 0.4372i	0.4212 − 0.2700i	0.6532 − 0.0565i
	−0.4917 − 0.1535i	−0.3293 − 0.2385i	0.3392 + 0.6740i
Binary 393 =	−0.3713 − 0.5263i	−0.7625 + 0.0549i	−0.0162 − 0.0227i
	−0.1727 − 0.5385i	0.4874 + 0.1129i	−0.2616 − 0.6012i
	−0.3957 + 0.3297i	−0.0592 − 0.4023i	0.5309 − 0.5362i
Binary 394 =	−0.9988 − 0.0498i	0	0
	0	0.2708 − 0.9626i	0
	0	0	−0.6617 + 0.7498i
Binary 395 =	−0.4930 + 0.5842i	0.0267 − 0.0081i	−0.1997 − 0.6123i
	0.4212 − 0.2700i	0.6532 − 0.0565i	−0.0063 − 0.5655i
	−0.3293 − 0.2385i	0.3392 + 0.6740i	−0.4753 + 0.1984i
Binary 396 =	0	0	0.3653 − 0.9309i
	0.9215 + 0.3884i	0	0
	0	−0.5837 + 0.8119i	0
Binary 397 =	0.3653 − 0.9309i	0	0
	0	0.9556 − 0.2948i	0
	0	0	0.0747 + 0.9972i
Binary 398 =	0.5216 + 0.4386i	−0.1753 + 0.5206i	−0.1755 + 0.4505i
	0.0958 − 0.1214i	−0.5215 − 0.5236i	0.3587 + 0.5489i
	0.3290 − 0.6351i	−0.3885 + 0.0356i	−0.5541 − 0.1707i
Binary 399 =	0	−0.3185 − 0.9479i	0
	0	0	0.5425 − 0.8400i
	0.0747 + 0.9972i	0	0
Binary 400 =	−0.2217 − 0.4297i	−0.6807 − 0.0347i	0.5466 − 0.0543i
	−0.6489 − 0.0942i	−0.0031 + 0.1546i	−0.2628 + 0.6907i
	0.4947 − 0.3025i	0.1212 + 0.7049i	0.1923 + 0.3394i
Binary 401 =	−0.9010 − 0.4339i	0	0
	0	0.6235 + 0.7818i	0
	0	0	−0.2225 + 0.9749i
Binary 402 =	0	−0.9010 + 0.4339i	0
	0	0	0.6235 − 0.7818i
	0.4562 − 0.8899i	0	0
Binary 403 =	0	−0.8533 + 0.5214i	0
	0	0	−0.1243 + 0.9922i
	0.8262 + 0.5633i	0	0
Binary 404 =	0.3536 − 0.3298i	0.4244 + 0.5333i	−0.5432 + 0.0814i
	−0.0850 − 0.6502i	0.1180 − 0.1000i	0.2280 − 0.7029i
	0.4252 + 0.3943i	0.4483 − 0.5573i	−0.2089 − 0.3294i
Binary 405 =	−0.6433 + 0.4129i	−0.0061 − 0.0272i	0.4970 − 0.4096i
	0.4820 − 0.1339i	−0.0075 − 0.6556i	0.5241 − 0.2124i
	−0.2443 − 0.3250i	0.6975 − 0.2879i	−0.3584 − 0.3700i
Binary 406 =	0	−0.0249 − 0.9997i	0
	0	0	−0.7971 − 0.6038i
	−0.9010 + 0.4339i	0	0
Binary 407 =	0.0589 − 0.4799i	0.6679 + 0.1357i	−0.3638 + 0.4116i
	−0.4831 − 0.4434i	0.0261 − 0.1524i	−0.2771 − 0.6850i
	0.5791 + 0.0287i	−0.0148 − 0.7151i	−0.3718 − 0.1181i
Binary 408 =	−0.4760 + 0.4339i	−0.7642 − 0.0212i	0.0020 + 0.0278i
	−0.5129 + 0.2383i	0.4737 + 0.1609i	−0.0903 + 0.6494i
	0.3764 + 0.3517i	−0.0188 − 0.4062i	−0.7326 + 0.1807i
Binary 409 =	−0.4030 + 0.3733i	0.0133 + 0.4833i	−0.3973 − 0.5538i
	−0.2076 − 0.7092i	0.5438 + 0.3665i	−0.1228 + 0.0940i
	−0.3582 − 0.1545i	−0.5770 + 0.0579i	−0.4755 + 0.5343i
Binary 410 =	0.7588 − 0.0929i	0.0175 − 0.0217i	0.2887 + 0.5757i
	−0.4924 − 0.0885i	0.5079 − 0.4146i	0.0905 + 0.5583i
	0.0791 + 0.3988i	0.6600 + 0.3658i	0.4404 − 0.2671i
Binary 411 =	−0.5778 − 0.2846i	−0.7239 − 0.2455i	0.0279 + 0.0001i
	−0.4092 − 0.3904i	0.4052 + 0.2934i	0.6408 + 0.1385i
	−0.1898 + 0.4788i	0.1017 − 0.3937i	0.1255 + 0.7441i
Binary 412 =	−0.6435 + 0.0264i	−0.5717 − 0.5075i	−0.0164 + 0.0226i
	−0.5461 − 0.1471i	0.2595 + 0.4278i	−0.4866 + 0.4394i
	0.0623 + 0.5113i	0.2467 − 0.3232i	−0.6774 − 0.3324i
Binary 413 =	−0.4515 + 0.5105i	−0.1745 − 0.5208i	0.4828 − 0.0254i
	0.1189 + 0.0988i	0.7318 + 0.1025i	0.3528 − 0.5527i
	0.6267 + 0.3447i	0.2881 − 0.2630i	0.0723 + 0.5753i
Binary 414 =	−0.6033 − 0.2256i	0.7458 − 0.1680i	−0.0036 + 0.0276i
	−0.4460 − 0.3477i	−0.4988 − 0.0390i	−0.2172 + 0.6186i
	−0.1412 + 0.4954i	0.1185 + 0.3890i	−0.7539 + 0.0320i
Binary 415 =	0	0	0.9556 − 0.2948i
	−0.5837 + 0.8119i	0	0
	0	0.8782 − 0.4783i	0
Binary 416 =	0	0	0.9950 − 0.0996i
	−0.9888 − 0.1490i	0	0
	0	−0.7331 + 0.6802i	0
Binary 417 =	0.4244 + 0.5333i	−0.5432 + 0.0814i	0.0589 − 0.4799i
	0.1180 − 0.1000i	0.2280 − 0.7029i	−0.4831 − 0.4434i
	0.4483 − 0.5573i	−0.2089 − 0.3294i	0.5791 + 0.0287i
Binary 418 =	0.0273 + 0.0056i	−0.4219 − 0.4867i	0.6220 − 0.4445i
	0.6007 + 0.2628i	−0.2254 − 0.5187i	−0.4748 + 0.1578i
	−0.0244 + 0.7542i	−0.3609 + 0.3675i	0.2602 + 0.3124i
Binary 419 =	0	0	0.9802 − 0.1981i
	0.9215 − 0.3884i	0	0
	0	−0.5837 − 0.8119i	0
Binary 420 =	−0.2274 + 0.7298i	−0.0222 − 0.0169i	0.1175 − 0.6333i
	0.2832 − 0.4124i	−0.4272 − 0.4974i	0.2649 − 0.4996i
	−0.3961 − 0.0919i	0.3492 − 0.6689i	−0.5124 − 0.0531i
Binary 421 =	0.4779 − 0.0733i	−0.3401 − 0.5906i	0.3428 − 0.4292i
	0.2960 − 0.5851i	−0.1316 + 0.0813i	0.3109 + 0.6704i
	0.1292 + 0.5653i	−0.5264 + 0.4843i	0.3772 + 0.0994i
Binary 422 =	0.5226 − 0.4375i	0.4822 + 0.2631i	0.4132 + 0.2510i
	−0.1029 − 0.1154i	−0.6062 + 0.4226i	0.6029 − 0.2580i
	−0.5683 − 0.4343i	−0.0324 + 0.3888i	−0.2644 + 0.5161i
Binary 423 =	0	−0.5000 − 0.8660i	0
	0	0	−0.5000 − 0.8660i
	0.1736 − 0.9848i	0	0
Binary 424 =	−0.4024 − 0.3739i	0.4548 + 0.1642i	−0.5428 − 0.4121i
	0.6917 − 0.2600i	0.5398 − 0.3723i	−0.0897 + 0.1260i
	0.1273 − 0.3687i	−0.1569 + 0.5582i	−0.2969 + 0.6507i
Binary 425 =	−0.0205 + 0.0189i	−0.6033 − 0.2256i	0.7458 − 0.1680i
	−0.5640 + 0.3343i	−0.4460 − 0.3477i	−0.4988 − 0.0390i
	−0.5981 − 0.4601i	−0.1412 + 0.4954i	0.1185 + 0.3890i
Binary 426 =	0.4970 − 0.4096i	−0.2274 + 0.7298i	−0.0222 − 0.0169i
	0.5241 − 0.2124i	0.2832 − 0.4124i	−0.4272 − 0.4974i
	−0.3584 − 0.3700i	−0.3961 − 0.0919i	0.3492 − 0.6689i
Binary 427 =	0.2635 − 0.7176i	−0.0223 + 0.0167i	0.3971 + 0.5071i
	−0.3034 + 0.3978i	−0.5945 + 0.2765i	0.1993 + 0.5293i
	0.3910 + 0.1115i	−0.5493 − 0.5173i	0.3788 − 0.3490i
Binary 428 =	0.6515 − 0.2002i	−0.4951 + 0.2379i	0.4830 + 0.0228i
	−0.0499 − 0.1463i	0.0107 − 0.7389i	0.4061 − 0.5149i
	−0.3550 − 0.6209i	−0.2968 − 0.2532i	0.0146 + 0.5797i
Binary 429 =	−0.0034 − 0.0277i	−0.6377 + 0.0903i	0.7524 + 0.1349i
	0.0578 − 0.6531i	−0.5580 − 0.0920i	−0.4444 − 0.2297i
	0.7227 − 0.2170i	0.1129 + 0.5026i	−0.0419 + 0.4044i
Binary 430 =	0.0267 − 0.0081i	−0.1997 − 0.6123i	−0.0022 + 0.7644i
	0.6532 − 0.0565i	−0.0063 − 0.5655i	0.1490 − 0.4776i
	0.3392 + 0.6740i	−0.4753 + 0.1984i	−0.4056 + 0.0289i
Binary 431 =	−0.5428 − 0.4121i	−0.5486 − 0.0278i	0.4539 − 0.1666i
	−0.0897 + 0.1260i	0.3628 − 0.6438i	0.1742 − 0.6322i
	−0.2969 + 0.6507i	−0.1395 − 0.3643i	0.2386 + 0.5285i
Binary 432 =	−0.1865 − 0.7413i	0.0195 − 0.0199i	0.0538 − 0.6418i
	−0.0266 + 0.4996i	0.5466 − 0.3620i	0.2139 − 0.5235i
	0.3859 − 0.1281i	0.6203 + 0.4297i	−0.5151 − 0.0018i
Binary 433 =	−0.9691 − 0.2468i	0	0
	0	0.9802 − 0.1981i	0
	0	0	0.4562 − 0.8899i
Binary 434 =	0.1736 − 0.9848i	0	0
	0	0.1736 − 0.9848i	0
	0	0	0.1736 − 0.9848i
Binary 435 =	−0.4598 − 0.6107i	−0.0275 − 0.0043i	0.3542 − 0.5379i
	0.1696 + 0.4707i	−0.6130 − 0.2325i	0.4382 − 0.3575i
	0.3058 − 0.2679i	−0.0132 − 0.7544i	−0.4515 − 0.2479i
Binary 436 =	−0.4002 − 0.2713i	−0.6137 − 0.2964i	−0.2750 + 0.4755i
	−0.6150 + 0.2276i	−0.0629 + 0.1412i	−0.4074 − 0.6165i
	0.2898 − 0.5023i	−0.1621 + 0.6966i	−0.3878 − 0.0420i
Binary 437 =	0.6793 + 0.3507i	−0.0205 + 0.0189i	−0.6033 − 0.2256i
	−0.3570 − 0.3505i	−0.5640 + 0.3343i	−0.4460 − 0.3477i
	−0.1593 + 0.3741i	−0.5981 − 0.4601i	−0.1412 + 0.4954i
Binary 438 =	−0.3638 + 0.4116i	−0.2634 − 0.4055i	0.5989 − 0.3253i
	−0.2771 − 0.6850i	−0.6551 − 0.0292i	−0.0780 − 0.1335i
	−0.3718 − 0.1181i	0.4621 − 0.3503i	−0.4710 − 0.5383i
Binary 439 =	0.9950 − 0.0996i	0	0
	0	−0.8533 + 0.5214i	0
	0	0	−0.1243 + 0.9922i
Binary 440 =	0	−0.7331 − 0.6802i	0
	0	0	0.8262 − 0.5633i
	−0.8533 − 0.5214i	0	0
Binary 441 =	−0.0222 − 0.0169i	0.1175 − 0.6333i	0.2949 + 0.7053i
	−0.4272 − 0.4974i	0.2649 − 0.4996i	−0.0482 − 0.4980i
	0.3492 − 0.6689i	−0.5124 − 0.0531i	−0.3625 + 0.1842i
Binary 442 =	−0.9888 + 0.1490i	0	0
	0	−0.7331 − 0.6802i	0
	0	0	0.8262 − 0.5633i
Binary 443 =	−0.3185 − 0.9479i	0	0
	0	0.5425 − 0.8400i	0
	0	0	0.6982 + 0.7159i
Binary 444 =	−0.4624 − 0.1413i	0.2497 − 0.6342i	0.3421 + 0.4297i
	−0.5206 + 0.3987i	−0.1456 − 0.0522i	−0.7228 + 0.1540i
	0.1289 − 0.5654i	−0.7068 − 0.1096i	−0.1808 + 0.3457i
Binary 445 =	0.5684 − 0.3030i	−0.7532 + 0.1306i	0.0256 + 0.0109i
	0.5558 − 0.1044i	0.4962 + 0.0638i	0.5367 + 0.3766i
	−0.2780 − 0.4337i	−0.0989 − 0.3944i	−0.1734 + 0.7344i
Binary 446 =	0.5465 + 0.0551i	−0.0374 − 0.4821i	0.4255 − 0.5324i
	−0.3944 + 0.6249i	−0.5614 − 0.3389i	−0.1237 − 0.0928i
	0.1212 + 0.3708i	0.5734 − 0.0866i	−0.6431 − 0.3131i
Binary 447 =	0.1689 + 0.6215i	0.7353 + 0.2091i	0.0074 + 0.0269i
	−0.0219 + 0.5651i	−0.4194 − 0.2728i	0.0402 + 0.6544i
	0.4846 − 0.1745i	−0.0820 + 0.3982i	−0.6823 + 0.3223i
Binary 448 =	−0.6807 + 0.0333i	−0.4381 + 0.3313i	0.2406 − 0.4194i
	0.0123 + 0.1541i	−0.1359 − 0.7264i	−0.2729 − 0.5962i
	0.1908 + 0.6893i	−0.3410 − 0.1894i	0.5225 + 0.2514i
Binary 449 =	0.0152 − 0.0234i	−0.5466 − 0.3407i	0.4897 + 0.5870i
	0.4641 − 0.4632i	−0.3683 − 0.4292i	−0.1928 − 0.4617i
	0.6931 + 0.2983i	−0.2366 + 0.4575i	−0.2921 + 0.2829i
Binary 450 =	−0.2797 − 0.6215i	−0.5163 + 0.1875i	−0.4002 − 0.2713i
	−0.1390 + 0.0678i	0.0842 − 0.7342i	−0.6150 + 0.2276i
	−0.5720 + 0.4295i	−0.2701 − 0.2815i	0.2898 − 0.5023i
Binary 451 =	−0.1118 − 0.7562i	−0.0270 + 0.0068i	−0.6435 + 0.0264i
	−0.0762 + 0.4945i	−0.6552 + 0.0239i	−0.5461 − 0.1471i
	0.3967 − 0.0891i	−0.3052 − 0.6901i	0.0623 + 0.5113i
Binary 452 =	0.4132 + 0.2510i	0.1191 − 0.6711i	0.5385 − 0.1084i
	0.6029 − 0.2580i	−0.1530 − 0.0223i	−0.1927 + 0.7134i
	−0.2644 + 0.5161i	−0.7145 + 0.0326i	0.2251 + 0.3186i
Binary 453 =	−0.2989 + 0.7036i	0.0101 + 0.0260i	−0.6430 − 0.0378i
	0.3228 − 0.3822i	0.1052 + 0.6472i	−0.5287 − 0.2008i
	−0.3850 − 0.1309i	−0.6468 + 0.3886i	0.0111 + 0.5150i
Binary 454 =	0.5385 − 0.1084i	0.4779 − 0.0733i	−0.3401 − 0.5906i
	−0.1927 + 0.7134i	0.2960 − 0.5851i	−0.1316 + 0.0813i
	0.2251 + 0.3186i	0.1292 + 0.5653i	−0.5264 + 0.4843i
Binary 455 =	0.5982 + 0.3266i	0.3832 + 0.3935i	0.2812 − 0.3933i
	0.0699 − 0.1379i	−0.7038 + 0.2252i	−0.2122 − 0.6205i
	0.1966 − 0.6877i	−0.1455 + 0.3619i	0.4949 + 0.3022i
Binary 456 =	−0.5990 + 0.4749i	0.0194 + 0.0200i	0.3446 + 0.5441i
	0.4663 − 0.1812i	0.3483 + 0.5555i	0.1456 + 0.5465i
	−0.2755 − 0.2990i	−0.4450 + 0.6094i	0.4117 − 0.3096i
Binary 457 =	−0.6256 + 0.1534i	−0.7124 + 0.2772i	0.0214 − 0.0178i
	−0.5644 − 0.0360i	0.4990 − 0.0357i	0.5800 − 0.3058i
	0.1624 + 0.4888i	−0.1751 − 0.3670i	0.5744 + 0.4893i
Binary 458 =	0	0	−0.0249 − 0.9997i
	−0.2225 − 0.9749i	0	0
	0	−0.9010 + 0.4339i	0
Binary 459 =	0.9215 − 0.3884i	0	0
	0	−0.5837 − 0.8119i	0
	0	0	0.8782 + 0.4783i
Binary 460 =	0	0	0.9215 − 0.3884i
	0.0747 − 0.9972i	0	0
	0	0.3653 + 0.9309i	0
Binary 461 =	−0.4682 − 0.4423i	−0.2989 + 0.7036i	0.0101 + 0.0260i
	−0.2760 − 0.4936i	0.3228 − 0.3822i	0.1052 + 0.6472i
	−0.3225 + 0.4016i	−0.3850 − 0.1309i	−0.6468 + 0.3886i
Binary 462 =	0.3446 + 0.5441i	−0.1536 + 0.7489i	0.0277 + 0.0029i
	0.1456 + 0.5465i	0.2407 − 0.4386i	0.6239 + 0.2017i
	0.4117 − 0.3096i	−0.4033 − 0.0520i	0.0508 + 0.7528i
Binary 463 =	−0.4951 + 0.2379i	0.4830 + 0.0228i	0.3704 − 0.5721i
	0.0107 − 0.7389i	0.4061 − 0.5149i	−0.1323 − 0.0800i
	−0.2968 − 0.2532i	0.0146 + 0.5797i	−0.6711 − 0.2475i
Binary 464 =	−0.4240 + 0.2324i	0.5216 + 0.4386i	−0.1753 + 0.5206i
	−0.0780 + 0.6511i	0.0958 − 0.1214i	−0.5215 − 0.5236i
	−0.3147 − 0.4870i	0.3290 − 0.6351i	−0.3885 + 0.0356i
Binary 465 =	−0.8533 − 0.5214i	0	0
	0	−0.1243 − 0.9922i	0
	0	0	0.9950 + 0.0996i
Binary 466 =	0.7643 − 0.0169i	−0.0262 + 0.0094i	0.5354 − 0.3581i
	−0.4812 − 0.1371i	−0.6496 + 0.0890i	0.5427 − 0.1592i
	0.0390 + 0.4047i	−0.3724 − 0.6563i	−0.3198 − 0.4038i
Binary 467 =	0.2812 − 0.3933i	0.6682 − 0.1343i	0.5465 + 0.0551i
	−0.2122 − 0.6205i	−0.0351 − 0.1506i	−0.3944 + 0.6249i
	0.4949 + 0.3022i	−0.2914 − 0.6532i	0.1212 + 0.3708i
Binary 468 =	0.5383 + 0.1093i	−0.2194 + 0.4308i	−0.5812 − 0.3560i
	−0.4547 + 0.5825i	0.3023 + 0.5819i	−0.0767 + 0.1343i
	0.0837 + 0.3810i	−0.5344 − 0.2251i	−0.2307 + 0.6770i
Binary 469 =	0	0	−0.4113 − 0.9115i
	−0.6617 + 0.7498i	0	0
	0	−0.9988 − 0.0498i	0
Binary 470 =	0.5457 + 0.5353i	0.0173 + 0.0219i	−0.0104 − 0.6440i
	−0.2378 − 0.4402i	0.2912 + 0.5874i	0.1607 − 0.5422i
	−0.2625 + 0.3105i	−0.5035 + 0.5620i	−0.5127 + 0.0495i
Binary 471 =	−0.6377 + 0.0903i	0.7524 + 0.1349i	−0.0204 − 0.0190i
	−0.5580 − 0.0920i	−0.4444 − 0.2297i	−0.3755 − 0.5374i
	0.1129 + 0.5026i	−0.0419 + 0.4044i	0.4141 − 0.6308i
Binary 472 =	0	−0.6617 − 0.7498i	0
	0	0	−0.9988 + 0.0498i
	−0.4113 + 0.9115i	0	0
Binary 473 =	−0.0008 + 0.0279i	0.4060 − 0.5000i	0.5457 + 0.5353i
	−0.1545 + 0.6372i	0.4716 − 0.3121i	−0.2378 − 0.4402i
	−0.7469 + 0.1069i	−0.4246 − 0.2916i	−0.2625 + 0.3105i
Binary 474 =	0	−0.7971 − 0.6038i	0
	0	0	−0.4113 + 0.9115i
	0.6235 − 0.7818i	0	0
Binary 475 =	0.6631 − 0.3803i	−0.0278 − 0.0015i	−0.0218 + 0.6437i
	−0.4881 + 0.1097i	−0.6331 − 0.1703i	−0.1875 + 0.5335i
	0.2278 + 0.3368i	−0.0882 − 0.7494i	0.5145 − 0.0239i
Binary 476 =	−0.5168 + 0.3843i	0.6410 + 0.4166i	0.0273 − 0.0054i
	−0.5341 + 0.1861i	−0.3203 − 0.3843i	0.6556 + 0.0088i
	0.3395 + 0.3874i	−0.1957 + 0.3564i	0.2704 + 0.7044i
Binary 477 =	0.1062 + 0.6353i	−0.5184 − 0.5619i	0.0257 − 0.0107i
	−0.0781 + 0.5601i	0.2156 + 0.4515i	0.6443 − 0.1213i
	0.4996 − 0.1254i	0.2776 − 0.2971i	0.4046 + 0.6369i
Binary 478 =	0.1484 + 0.5289i	0.1063 − 0.4717i	−0.4991 − 0.4641i
	−0.7258 − 0.1389i	−0.4365 − 0.4893i	−0.1018 + 0.1164i
	−0.3009 + 0.2483i	0.5734 + 0.0862i	−0.3602 + 0.6179i
Binary 479 =	−0.1226 + 0.5354i	−0.0852 − 0.4759i	−0.0855 + 0.6762i
	−0.5710 − 0.4691i	−0.5923 − 0.2813i	0.1539 + 0.0147i
	−0.3830 + 0.0741i	0.5619 − 0.1433i	0.7120 − 0.0682i
Binary 480 =	−0.9010 + 0.4339i	0	0
	0	0.6235 − 0.7818i	0
	0	0	−0.2225 − 0.9749i
Binary 481 =	0.0173 + 0.0219i	−0.0104 − 0.6440i	0.7621 + 0.0593i
	0.2912 + 0.5874i	0.1607 − 0.5422i	−0.4651 − 0.1843i
	−0.5035 + 0.5620i	−0.5127 + 0.0495i	−0.0015 + 0.4066i
Binary 482 =	0.3704 − 0.5721i	−0.2263 + 0.5005i	0.3846 − 0.2930i
	−0.1323 − 0.0800i	−0.4668 − 0.5729i	−0.0199 − 0.6554i
	−0.6711 − 0.2475i	−0.3901 − 0.0032i	0.3838 + 0.4346i
Binary 483 =	−0.5432 + 0.0814i	0.0589 − 0.4799i	0.6679 + 0.1357i
	0.2280 − 0.7029i	−0.4831 − 0.4434i	0.0261 − 0.1524i
	−0.2089 − 0.3294i	0.5791 + 0.0287i	−0.0148 − 0.7151i
Binary 484 =	0.3846 − 0.2930i	−0.0840 − 0.6763i	0.1484 + 0.5289i
	−0.0199 − 0.6554i	−0.1528 + 0.0238i	−0.7258 − 0.1389i
	0.3838 + 0.4346i	−0.6731 + 0.2418i	−0.3009 + 0.2483i
Binary 485 =	0.1063 − 0.4717i	−0.4991 − 0.4641i	0.4536 + 0.3098i
	−0.4365 − 0.4893i	−0.1018 + 0.1164i	−0.6453 + 0.3602i
	0.5734 + 0.0862i	−0.3602 + 0.6179i	−0.0709 + 0.3836i
Binary 486 =	0.6220 − 0.4445i	0.0245 − 0.0132i	−0.6360 − 0.1016i
	−0.4748 + 0.1578i	0.6291 − 0.1848i	−0.5061 − 0.2524i
	0.2602 + 0.3124i	0.4661 + 0.5934i	−0.0403 + 0.5135i
Binary 487 =	0.4060 − 0.5000i	0.5457 + 0.5353i	0.0173 + 0.0219i
	0.4716 − 0.3121i	−0.2378 − 0.4402i	0.2912 + 0.5874i
	−0.4246 − 0.2916i	−0.2625 + 0.3105i	−0.5035 + 0.5620i
Binary 488 =	−0.3988 + 0.2734i	−0.6738 − 0.1023i	−0.4947 − 0.2387i
	−0.0128 + 0.6556i	−0.0185 + 0.1535i	0.5844 − 0.4523i
	−0.3617 − 0.4532i	0.0504 + 0.7135i	0.0129 − 0.3899i
Binary 489 =	0.6679 + 0.1357i	−0.3638 + 0.4116i	−0.2634 − 0.4055i
	0.0261 − 0.1524i	−0.2771 − 0.6850i	−0.6551 − 0.0292i
	−0.0148 − 0.7151i	−0.3718 − 0.1181i	0.4621 − 0.3503i
Binary 490 =	0.0101 + 0.0260i	−0.6430 − 0.0378i	0.2232 + 0.7311i
	0.1052 + 0.6472i	−0.5287 − 0.2008i	0.0016 − 0.5003i
	−0.6468 + 0.3886i	0.0111 + 0.5150i	−0.3790 + 0.1472i
Binary 491 =	0.7660 − 0.6428i	0	0
	0	0.7660 − 0.6428i	0
	0	0	0.7660 − 0.6428i
Binary 492 =	−0.0114 − 0.0254i	−0.1378 − 0.6292i	0.4005 − 0.6512i
	−0.1373 − 0.6411i	0.0501 − 0.5633i	−0.3762 + 0.3298i
	0.6266 − 0.4204i	−0.4927 + 0.1501i	0.3612 + 0.1868i
Binary 493 =	−0.5001 + 0.4630i	−0.1226 + 0.5354i	−0.0852 − 0.4759i
	0.1085 + 0.1102i	−0.5710 − 0.4691i	−0.5923 − 0.2813i
	0.5893 + 0.4054i	−0.3830 + 0.0741i	0.5619 − 0.1433i
Binary 494 =	−0.1745 − 0.5208i	0.4828 − 0.0254i	−0.0177 + 0.6813i
	0.7318 + 0.1025i	0.3528 − 0.5527i	0.1546 − 0.0007i
	0.2881 − 0.2630i	0.0723 + 0.5753i	0.7017 − 0.1388i
Binary 495 =	0.6235 − 0.7818i	0	0
	0	−0.2225 − 0.9749i	0
	0	0	−0.9010 + 0.4339i
Binary 496 =	−0.0141 + 0.5491i	−0.4624 − 0.1413i	0.2497 − 0.6342i
	−0.6526 − 0.3466i	−0.5206 + 0.3987i	−0.1456 − 0.0522i
	−0.3607 + 0.1486i	0.1289 − 0.5654i	−0.7068 − 0.1096i
Binary 497 =	−0.4835 + 0.0013i	0.6511 + 0.2016i	0.5250 − 0.1615i
	−0.3799 + 0.5345i	0.0412 − 0.1490i	−0.1207 + 0.7291i
	−0.0435 − 0.5782i	0.0565 − 0.7130i	0.2557 + 0.2946i
Binary 498 =	−0.1778 − 0.4496i	0.5627 + 0.3845i	−0.1218 − 0.5356i
	−0.6363 − 0.1584i	0.0833 − 0.1303i	0.7180 + 0.1749i
	0.5224 − 0.2517i	0.2641 − 0.6647i	0.3129 − 0.2330i
Binary 499 =	0	0.9215 − 0.3884i	0
	0	0	−0.5837 − 0.8119i
	0.3653 + 0.9309i	0	0
Binary 500 =	0	0	0.8262 − 0.5633i
	−0.8533 − 0.5214i	0	0
	0	−0.1243 − 0.9922i	0
Binary 501 =	0.0077 − 0.0268i	0.2989 − 0.5705i	0.7253 − 0.2414i
	0.3069 − 0.5794i	0.4004 − 0.3993i	−0.5002 + 0.0108i
	0.7502 + 0.0807i	−0.4739 − 0.2017i	0.1566 + 0.3752i
Binary 502 =	0	0.4562 − 0.8899i	0
	0	0	−0.9691 − 0.2468i
	0.8782 + 0.4783i	0	0
Binary 503 =	−0.2743 − 0.4759i	0.1977 − 0.4412i	0.1853 − 0.6559i
	0.7376 − 0.0445i	−0.3309 − 0.5661i	−0.1500 − 0.0374i
	0.2303 − 0.3148i	0.5449 + 0.1982i	−0.7142 − 0.0387i
Binary 504 =	−0.2105 + 0.6087i	−0.6959 − 0.3164i	−0.0237 − 0.0146i
	−0.3365 + 0.4546i	0.3740 + 0.3323i	−0.4746 − 0.4524i
	0.4987 + 0.1288i	0.1404 − 0.3816i	0.2809 − 0.7003i
Binary 505 =	−0.0177 + 0.6813i	−0.4685 − 0.2868i	0.3536 − 0.3298i
	0.1546 − 0.0007i	0.6265 − 0.3919i	−0.0850 − 0.6502i
	0.7017 − 0.1388i	0.0517 − 0.3867i	0.4252 + 0.3943i
Binary 506 =	−0.3171 − 0.5606i	0.6793 + 0.3507i	−0.0205 + 0.0189i
	−0.1182 − 0.5530i	−0.3570 − 0.3505i	−0.5640 + 0.3343i
	−0.4266 + 0.2887i	−0.1593 + 0.3741i	−0.5981 − 0.4601i
Binary 507 =	0.5060 + 0.2138i	−0.4835 + 0.0013i	0.6511 + 0.2016i
	−0.5611 + 0.4809i	−0.3799 + 0.5345i	0.0412 − 0.1490i
	0.0065 + 0.3900i	−0.0435 − 0.5782i	0.0565 − 0.7130i
Binary 508 =	−0.6072 + 0.2149i	0.4633 − 0.6080i	0.0150 + 0.0235i
	−0.5652 + 0.0204i	−0.4072 + 0.2907i	0.2313 + 0.6135i
	0.2102 + 0.4702i	0.3408 + 0.2218i	−0.5569 + 0.5091i
Binary 509 =	−0.5486 − 0.0278i	0.4539 − 0.1666i	−0.6807 + 0.0333i
	0.3628 − 0.6438i	0.1742 − 0.6322i	0.0123 + 0.1541i
	−0.1395 − 0.3643i	0.2386 + 0.5285i	0.1908 + 0.6893i
Binary 510 =	0	0	−0.9888 − 0.1490i
	−0.9988 + 0.0498i	0	0
	0	0.2708 + 0.9626i	0
Binary 511 =	0	0.7660 − 0.6428i	0
	0	0	0.7660 − 0.6428i
	1.0000 − 0.0000i	0	0

Claims

1. A method of transmitting data, comprising the steps of: (a) providing a set of at least 2m n×n matrices that represent an extension of a fixed-point-free group, each said matrix including n2 matrix elements, where m is a positive integer and n is an integer greater than 1;(b) allocating each binary number between 0 and binary 2′-1 to a respective one of said matrices;(c) mapping the data into said matrices according to said allocating; and(d) for each said matrix into which the data are mapped, transmitting said matrix elements of said each matrix.

2. The method of claim 1, wherein said transmitting includes, for each said matrix into which the data are mapped, successively for each column of said each matrix, transmitting each said matrix element of said each column via a different respective antenna.

3. The method of claim 1, wherein said transmitting includes, for each said matrix into which the data are mapped, transmitting each said matrix element of each column of said each matrix via a different respective antenna and transmitting each said matrix element of each row of said each matrix at a different respective frequency.

4. The method of claim 1, wherein said transmitting includes, for each said matrix into which the data are mapped, successively for each column of said each matrix, transmitting each said matrix element of said each column at a different respective frequency.

5. The method of claim 1, wherein said fixed-point-free group is a quaternion group.

6. The method of claim 5, wherein said extension is a super quaternion set.

7. The method of claim 1, wherein said fixed-point-free group is a Gm,r group.

8. The method of claim 1, wherein said extension is a union of said fixed-point-free group and at least one coset determined by an element of an algebra in which said fixed-point-free group resides.

9. The method of claim 1, further comprising the step of: (e) receiving said matrix elements of said each matrix via a known channel.

10. The method of claim 1, further comprising the step of: (f) receiving said matrix elements of said each matrix via an unknown channel.

11. A transmitter for transmitting data, comprising: (a) a coder for mapping the data into a set of 2m n×n matrices obtained by providing a set of at least 2m n×n matrices that represent an extension of a fixed-point-free group and allocating each binary number between 0 and binary 2m−1 to a respective one of said at least 2m matrices, each said matrix into which the data are mapped including n2 matrix elements, m being a positive integer, n being an integer greater than 1, said mapping being according to said allocating; and(b) at least one antenna for transmitting said matrix elements.

12. The transmitter of claim 11, comprising n said antennas, each said antenna for transmitting said matrix elements of a respective row of each said matrix into which the data are mapped.