Wireless communication system having linear encoder

Abstract

In general, linear complex-field encoding techniques are proposed For example, transmitter of a wireless communication system includes an encoder and a modulator. The encoder linearly encodes a data stream to produce an encoded data stream. The modulator to produce an output waveform in accordance with the encoded data stream for transmission through a wireless channel. The modulator generates the output waveform as a multicarrier waveform having a set of subcarriers, e.g., an Orthogonal Frequency Division Multiplexing (OFDM) waveform. The encoder linearly encodes the data stream so that the subcarriers carry different linear combinations of information symbols of the data stream.

Description

TECHNICAL FIELD

The invention relates to communication systems and, more particularly, transmitters and receivers for use in wireless communication systems.

BACKGROUND

In wireless mobile communications, a channel that couples a transmitter to a receiver is often time-varying due to relative transmitter-receiver motion and multipath propagation. Such a time-variation is commonly referred to as fading, and may severely impair system performance. When a data rate for the system is high in relation to channel bandwidth, multipath propagation may become frequency-selective and cause intersymbol interference (ISI). By implementing Inverse Fast Fourier Transform (IFFT) at the transmitter and FFT at the receiver, Orthogonal Frequency Division Multiplexing (OFDM) converts an ISI channel into a set of parallel ISI-free subchannels with gains equal to the channel's frequency response values on the FFT grid. Each subchannel can be easily equalized by a single-tap equalizer using scalar division.

To avoid inter-block interference (IBI) between successive IFFT processed blocks, a cyclic prefix (CP) of length greater than or equal to the channel order is inserted per block at the transmitter and discarded at the receiver. In addition to suppressing IBI, the CP also converts linear convolution into cyclic convolution and thus facilitates diagonalization of an associated channel matrix.

Instead of having multipath diversity in the form of (superimposed) delayed and scaled replicas of the transmitted symbols as in the case of serial transmission, OFDM transfers the multipath diversity to the frequency domain in the form of (usually correlated) fading frequency response. Each OFDM subchannel has its gain being expressed as a linear combination of the dispersive channel taps. When the channel has nulls (deep fades) close to or on the FFT grid, reliable detection of the symbols carried by these faded subcarriers becomes difficult if not impossible.

Error-control codes are usually invoked before the IFFT processing to deal with the frequency-selective fading. These include convolutional codes, Trellis Coded Modulation (TCM) or coset codes, Turbo-codes, and block codes (e.g., Reed-Solomon or BCH). Such coded OFDM schemes often incur high complexity and/or large decoding delay. Some of these schemes also require Channel State Information (CSI) at the transmitter, which may be unrealistic or too costly to acquire in wireless applications where the channel is rapidly changing. Another approach to guaranteeing symbol detectability over ISI channels is to modify the OFDM setup: instead of introducing the CP, each IFFT-processed block can be zero padded (ZP) by at least as many zeros as the channel order.

SUMMARY

In general, techniques are described for robustifying multi-carrier wireless transmissions, e.g., OFDM, against random frequency-selective fading by introducing memory into the transmission with complex field (CF) encoding across the subcarriers. Specifically, instead of sending a different uncoded symbol per subcarrier, the techniques utilize different linear combinations of the information symbols on the subcarriers. These techniques generalize signal space diversity concepts to allow for redundant encoding. The CF block code described herein can also be viewed as a form of real-number or analog codes.

The encoder described herein is referred to as a “Linear Encoder (LE),”0 and the corresponding encoding process is called “linear encoding,” also abbreviated as LE when no confusions arise. The resulting CF coded OFDM will be called LE-OFDM. In one embodiment, the linear encoder is designed so that maximum diversity order can be guaranteed without an essential decrease in transmission rate.

By performing pairwise error probability analysis, we upper bound the diversity order of OFDM transmissions over random frequency-selective fading channels. The diversity order is directly related to a Hamming distance between the coded symbols. Moreover, the described LE can be designed to guarantee maximum diversity order irrespective of the information symbol constellation with minimum redundancy. In addition, the described LE codes are maximum distance separable (MDS) in the real or complex field, which generalizes the well-known MDS concept for Galois field (GF) codes. Two classes of LE codes are described that can achieve MDS and guarantee maximum diversity order: the Vandermonde class, which generalizes the Reed-Solomon codes to the real/complex field, and the Cosine class, which does not have a GF counterpart.

Several possible decoding options have been described, including ML, ZF, MMSE, DFE, and iterative detectors. Decision directed detectors may be used to strike a trade-off between complexity and performance.

In one embodiment, a wireless communication device comprises an encoder that linearly encodes a data stream to produce an encoded data stream, and a modulator to produce an output waveform in accordance with the encoded data stream for transmission through a wireless channel.

In another embodiment, a wireless communication device comprises a demodulator that receives a waveform carrying a linearly encoded transmission and produces a demodulated data stream, and a decoder that applies decodes the demodulated data and produce estimated data.

In another embodiment, a method comprises linearly encoded a data stream with to produce an encoded data stream, and outputting a waveform in accordance with the data stream for transmission through a wireless channel.

In another embodiment, a computer-readable medium comprises instructions to cause a programmable processor to linearly encode a data stream with to produce an encoded data stream, and output a waveform in accordance with the data stream for transmission through a wireless channel.

The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the invention will be apparent from the description and drawings, and from the claims.

BRIEF DESCRIPTION OF DRAWINGS

FIG. 1 is a block diagram illustrating an exemplary wireless communication system in which a transmitter and receiver implement linear precoding techniques.

FIGS. 2A, 2B illustrate uncoded and GF-coded BPSK signals.

FIG. 3 illustrates an example format of a transmission block for CP-only transmissions by the transmitter of FIG. 1.

FIG. 4 illustrates an example format of a transmission block for ZP-only transmissions by the transmitter of FIG. 1.

FIG. 5 illustrates sphere decoding applied in one embodiment of the receiver of FIG. 1.

FIG. 6 illustrates an example portion of the receiver of FIG. 1.

FIG. 7 is factor graph representing an example linear encoding process.

FIGS. 8-10 are graphs that illustrate exemplary results of simulations of the described techniques.

DETAILED DESCRIPTION

FIG. 1 is a block diagram illustrating a telecommunication system 2 in which transmitter 4 communicates data to receiver 6 through wireless channel 8. Transmitter 4 transmits data to receiver 6 using one of a number of conventional multi-carrier transmission formats including Orthogonal Frequency Division Multiplexing (OFDM). OFDM has been adopted by many standards including digital audio and video broadcasting (DAB, DVB) in Europe and high-speed digital subscriber lines (DSL) in the United States. OFDM has also been proposed for local area mobile wireless broadband standards including IEEE802.11a, MMAC and HIPERLAN/2. In one embodiment, system 2 represents an LE-OFDM system having N subchannels.

In general, the techniques described herein robustify multi-carrier wireless transmissions, e.g., OFDM, against random frequency-selective fading by introducing memory into the transmission with complex field (CF) encoding across the subcarriers. In particular, transmitter 4 utilizes different linear combinations of the information symbols on the subcarriers. The techniques described herein may be applied to uplink and/or downlink transmissions, i.e., transmissions from a base station to a mobile device and vice versa. Consequently, transmitters 4 and receivers 6 may be any device configured to communicate using a multi-user wireless transmission including a cellular distribution station, a hub for a wireless local area network, a cellular phone, a laptop or handheld computing device, a personal digital assistant (PDA), and the like.

In the illustrated embodiment, transmitter 4 includes linear encoder 10 and an OFDM modulator 12. Receiver 6 includes OFDM demodulator 14 and equalizer 16. Due to CP-insertion at transmitter 44 and CP-removal at receiver 6, the dispersive channel 8 is represented as an N×N circulant matrix {tilde over (H)}, with [{tilde over (H)}]_i,j=h((i−j)mod N), where h(·) denotes the impulse response of channel 8:

$\begin{array}{cc}H=[\begin{array}{cccccccc}h\left(0\right)& 0& \dots & & 0& h\left(L\right)& \dots & h\left(1\right)\\ ⋮& h\left(0\right)& 0& \dots & & ⋰& ⋰& ⋮\\ & & & & & & & h\left(L\right)\\ h\left(L\right)& ⋮& ⋰& ⋰& & \dots & 0& 0\\ 0& h\left(L\right)& & ⋰& 0& & & 0\\ ⋮& 0& & & h\left(0\right)& ⋰& & ⋮\\ & ⋮& ⋰& ⋰& ⋮& ⋰& & 0\\ 0& \dots & & 0& h\left(L\right)& & \dots & h\left(0\right)\end{array}]& \left(1\right)\end{array}$ We assume the channel to be random FIR, consisting of no more than L+1 taps. The blocks within the dotted box represent a conventional uncoded OFDM system.

Let F denote the N×N FFT matrix with entries [F]_n,k=(1/√{square root over (N)})exp(−j2πnk/N). Performing IFFT (postmultiplication with the matrix F^H) at the transmitter and FFT (premultiplication with the matrix F) at the receiver diagonalizes the circulant matrix {tilde over (H)}. So, we obtain the parallel ISI-free model for the ith OFDM symbol as (see FIG. 1): x_i=D_Hμ_i+η_i, where

$D_H:=\mathrm{diag}\left[H\left(j0\right).H\left(j2\pi \frac{1}{N}\right),\dots ,H\left(j2\pi \frac{N-1}{N}\right)\right]=F^{}\tilde{H}F,$ with H(jω) denoting the channel frequency response at ω; and η_i=Fη_i standing for the FFT-processed additive white Gaussian noise (AWGN).

In order to exploit the frequency-domain diversity in OFDM, our LE-OFDM design first linearly encodes (i.e., maps) the K≦N symbols of the ith block, s_iε, where is the set of all possible vectors that s_i may belong to (e.g., the BPSK set {±1}^K×1), by an N×K matrix ΘεC^N×K and then multiplexes the coded symbols u_i=Θs_iεC^N×1 using conventional OFDM. In practice, the set is always finite. But we allow it to be infinite in our performance analysis. The encoder Θ considered here does not depend on the OFDM symbol index i. Time-varying encoder may be useful for certain purposes (e.g., power loading), but they will not be pursued here. Hence, from now on, we will drop our OFDM symbol index i for brevity.

Notice that the matrix-vector multiplication used in defining u=Θs takes place in the complex field, rather than a Galois field. The matrix Θ can be naturally viewed as the generating matrix of a complex field block code. The codebook is defined as :={Θs|sε}. By encoding a length-K vector to a length-N vector, some redundancy is introduced that we quantify by the rate of the code defined to be r=K/N, reminiscent of the GF block code rate definition. The set is a subset of the C^N×1 vector space. More specifically, is a subset of the K dimensional subspace spanned by the columns of Θ. When =Z^K×1, the set forms a lattice.

Combining the encoder with the diagonalized channel model, the ith received block after CP removal and FFT processing can be written as:x=F{tilde over (x)}=F({tilde over (H)}F^HΘs+ñ)=D_HΘs+η.  (2)We want to design Θ so that a large diversity order can be guaranteed irrespective of the constellation that the entries of s_i are drawn from, with a small amount of introduced redundancy.

We can conceptually view Θ together with the OFDM modulation F^H as a combined N×K encoder {tilde over (Θ)}:=F^HΘ, which in a sense blends the single-carrier and multicarrier notions. Indeed, by selecting Θ, hence {tilde over (Θ)}, the system in FIG. 1 can describe various single and multicarrier systems, some of them are provided shortly as special cases of our LE-OFDM. The received vector {tilde over (x)} is related to the information symbol vector s through the matrix product {tilde over (H)}Θ.

We define the Hamming distance δ(u, u′) between two vectors u and u′ as the number of non-zero entries in the vector u_c=u−u′ and the minimum Hamming distance of the set as δ_min ():=min{δ(u, u′)|u, uε}. When there is no confusion, we will simply use δ_min for brevity. The minimum Euclidean distance between vectors in is denoted as d_min() or simply d_min.

Because such encoding operates in the complex field, it does not increase the dimensionality of the signal space. This is to the contrasted to the GF encoding: the codeword set of a GF (n, k) code, when viewed as a real/complex vector, in general has a higher dimensionality (n) than does the original uncoded block of symbols (k). Exceptions include the repetition code, for which the codeword set has the same dimensionality as that of the input.

EXAMPLE 1

Consider the binary (3, 2) block code generated by the matrix

$\begin{array}{cc}{[\begin{array}{ccc}1& 0& 1\\ 0& 1& 1\end{array}]}^T& \left(3\right)\end{array}$ followed by BPSK constellation mapping (e.g., 0→−1 and 1→1). The codebook consists of 4 codewords[−1 −1 −1]^T, [1 −1 1]^T, [−1 1 1]^T, [1 1 −1]^T.  (4)These codewords span the R^3×1 (or C^3×1) space and therefore the codebook has dimension 3 in the real or complex field, as illustrated in FIG. 2.

In general, a (n, k) binary GF block code is capable of generating 2^k codewords in an n-dimensional space R^n×1 or C^n×1. If we view the transmit signal design problem as packing spheres in the signal space (Shannon's point of view), an (n, k) GF block code followed by constellation mapping packs spheres in an n-dimensional space and thus has the potential to be better (larger sphere radius) than a k-dimensional packing. In our example above, if we normalize the codewords by a factor √{square root over (2/3)} so that the energy per bit E_b is one, the 4 codewords have mutual Euclidean distance √{square root over (8/3)}, larger than the minimum distance √{square root over (2)} of the uncoded BPSK signal set (±1,±1). This increase in minimum Euclidean distance leads to improved system performance in AWGN channels, at least for high signal to noise ratio (SNR). For fading channels, the minimum Hamming distance of the codebook dominates high SNR performance in the form of diversity gain (as will become clear later). The diversity gain achieved by the (3, 2) block code in the example is the minimum Hamming distance 2.

CF linear encoding on the other hand, does not increase signal dimension; i.e., we always have dim(U)≦dim(S). When Θ has full column rank K, dim(U)=dim(S), in which case the codewords span a K-dimensional subspace of the N-dimensional vector space C^K×1. In terms of sphere packing, CF linear encoding does not yield a packing of dimension higher than K.

We have the following assertion about the minimum Euclidean distance.

Proposition 1 Suppose tr(ΘΘ^H)=K. If the entries of sε are drawn independently from a constellation of minimum Euclidean distance of d_min () then the codewords in u:={Θs|sε} have minimum Euclidean distance no more than d_min().

Proof: Under the power constraint tr(ΘΘ^H)=K, at least one column of Θ will have norm no more than 1. Without loss of generality, suppose the first column has norm no more than 1. Consider s_α=(α, 0, . . . , 0)^T and s_β=(β, 0, . . . , 0)^T, where α and β are two symbols from the constellation that are separated by d_min. The coded vectors u_α=Θs_α and u_β=Θs_β are then separated by a distance no more than d_min.

Due to Proposition 1, CF linear codes are not effective for improving performance for AWGN channels. But for fading channels, they may have an advantage over GF codes, because they are capable of producing codewords that have large Hamming distance.

EXAMPLE 2

The encoder

$\begin{array}{cc}\Theta ={\sqrt{\frac{4}{15}}[\begin{array}{ccc}1& 1& 1\\ 0.5& -0.5& 0.5\end{array}]}^T,& \left(5\right)\end{array}$ operating on BPSK signal set ={±1}², produces 4 codewords of minimum Euclidean distance √{square root over (4/5)} and minimum Hamming distance 3. Compared with the GF code in Example 1, this real code has smaller Euclidean distance but larger Hamming distance. In addition, the CF coding scheme described herein differs from the GF block coding in that the entries of the LE output vector u usually belong to a larger, although still finite, alphabet set than do the entries of the input vector s.

Before exploring optimal designs of Θ, let us first look at some special cases of the LE-OFDM system.

By setting K=N and Θ=I_N, we obtain the conventional uncoded OFDM model. In such a case, the one-tap linear equalizer matrix Γ=D_H⁻¹ yields ŝ=Γx=s+D_H⁻¹η, where the inverse exists when the channel has no nulls on the FFT grid. Under the assumption that ñ (hence η) is AWGN, such an equalizer followed by a minimum distance quantizer is optimum in the maximum-likelihood (ML) sense for a given channel when CSI has been acquired at the receiver. But when the channel has nulls on (or close to) the FFT grid ω−2πn/N, n=0, . . . , N−1, the matrix D_H will be ill-conditioned and serious noise-amplification will emerge if we try to invert D_H (the noise variance can become unbounded). Although events of channel nulls being close to the FFT grid have relatively low probability, their occurrence is known to have dominant impact on the average system performance especially at high SNR. Improving the performance of an uncoded transmission thus relies on robustifying the system against the occurrence of such low-probability but catastrophic events. If CSI is available at the transmitter, power and bit loading can be used and channel nulls can be avoided, such as in discrete multi-tone (DMT) systems.

If we choose K=N and Θ=F, then since F^HF=I_N, the IFFT F^H reverses the encoding and the resulting system is a single-carrier block transmission with CP insertion (c.f., FIG. 3): {tilde over (x)}={tilde over (H)}s+{tilde over (η)}. The FFT at the receiver is no longer necessary.

Let K=N−L. We choose Θ to be an N×K truncated FFT matrix (the first K columns of F); i.e., [Θ]_n,k=(1/√{square root over (N)}) exp(−j2πnk/N). It can be easily verified that F^HΘ=[I_K,0_K×L]^T:=T_zp, where 0_K×L denotes a K×L all-zero matrix, and the subscript “zp” stands for zero-padding (ZP). The matrix T_zp simply pads zeros at the tail of s and the zero-padded block ũ=T_zps is transmitted. Notice that H:={tilde over (H)}F^HΘ={tilde over (H)}T_zp is an N×K Toeplitz convolution matrix (the first K columns of {tilde over (H)}), which is always full rank. The symbols s can thus always be recovered from the received signal {tilde over (x)}=Hs+ñ (perfectly in the absence of noise) and no catastrophic channels exist in this case. The cyclic prefix in this case consists of L zeros, which, together with the L zeros from the encoding process, result in 2L consecutive zeros between two consecutive uncoded information blocks of length K. But only L zeros are needed in order to separate the information blocks. CP is therefore not necessary because the L zeros created by Θ already separate successive blocks.

ZP-only transmission is essentially a simple single-carrier block scheme. However, viewing it as a special case of the LE-OFDM design will allow us to apply the results about LE-OFDM and gain insights into its performance. It turns out that this special case is indeed very special: it achieves the best high-SNR performance among the LE-OFDM class.

To design linear encoder 10 with the goal of improving performance over uncoded OFDM, we utilize pair-wise error probability (PEP) analysis technique. For simplicity, we will first assume that

• As1) The channel h:=[h(0), h(1), . . . , h(L)]^T has independent and identically distributed (i.i.d.) zero-mean complex Gaussian taps (Rayleigh fading). The corresponding correlation matrix of h is R_h:=E[hh^H]=α_L²I_L+1, where the constant α_L:=1/(L+1).

Later on, we will relax this assumption to allow for correlated fading with possibly rank deficient autocorrelation matrix R_h.

We suppose ML detection with perfect CSI at the receiver and consider the probability P(s→s′|h), s, s′ε that a vector s is transmitted but is erroneously decoded as s′≠s. We define the set of all possible error vectors _e:={e:=s−s′|s.s′ε, s≠s′}.

The PEP can be approximated using the Chernoff bound as:P(s→s′|h)≦exp(−d²(y,y′)/4N₀),  (6)where N₀/2 is the noise variance per dimension, y:=D_HΘs, y′:=D_HΘs′, and d(y,y′)=∥y−y′∥ is the Euclidean distance between y and y′.

Let us consider the N×(L+1) matrix V with entries [V]_n,t=exp(−j2πnl/N), and use it to perform the N-point discrete Fourier transform Vh of h. Note that D_H=diag(Vh); i.e., the diagonal entries of D_H are those in vector Vh. Using the definitions e:=s−s′ε_e, u_e:=Θe, and D_e:=diag(u_e), we can write y−y′=D_Hu_e=diag(Vh)u_e. Furthermore, we can express the squared Euclidean distance d²(y,y′)=∥D_Hu_e∥²=∥D_eVh∥² asd²(y,y′)=h^HV^HD_e^HD_eVh:=h^HA_eh.  (7)

An upper bound to the average PEP can be obtained by averaging (6) with respect to the random channel h to obtain:

$\begin{array}{cc}P\left(s\to s^{\prime }\right)\le \prod _{l=0}^L\frac{1}{1+{\alpha }_L{\lambda }_{e,l}/\left(4N_0\right)},& \left(8\right)\end{array}$ where λ_e,0, λ_e,1, . . . , λ_e,L are the non-increasing eigen-values of the matrix A_e=V^HD_e^HD_eV.

If r_e is the rank of A_e, then λ_e,1≠0 if and only if 1ε[0,r_e−1]. Since 1+α_Lλ_e,1/(4N₀)>λ_e,1/(4N₀), it follows from (8) that

$\begin{array}{cc}P\left(s\to s^{\prime }\right)\le {\left(\frac{1}{4N_0}\right)}^{-r_e}{\left(\prod _{l=0}^{r_e-1}{\alpha }_L{\lambda }_{e,l}\right)}^{-1}.& \left(9\right)\end{array}$ We call r_e the diversity order, denoted as G_d,e, and (Π_t=0^re-1α_Lλ_e,l)^1/re the coding advantage, denoted as G_c,e, for the symbol error vector e. The diversity order G_d,e determines the slope of the averaged (w.r.t. the random channel) PEP (between s and s′) as a function of the SNR at high SNR (N₀→0). Correspondingly, G_c,e determines the shift of this PEP curve in SNR relative to a benchmark error rate curve of (¼N₀)^−re. When r_e=L+1, A_e is full rank, the product of eigen-values becomes the determinant of A_e and therefore the coding advantage is given by α_L[det(A_e)]^1/(L+1).

Since both G_d,e and G_d,c depend on the choice of e, we define the diversity order and coding advantages for our LE-OFDM system, respectively, as:

$\begin{array}{cc}{G}_d:=\underset{e\in e_{}}{\min }{G}_{d,e}=\underset{e\in e_{}}{\min }\mathrm{rank}\left(A_e\right),\mathrm{and}{G}_c:=\underset{e\in e_{}}{\min }{G}_{c,e}.& \left(10\right)\end{array}$

We refer to diversity order herein to mean the asymptotic slope of the error probability versus SNR curve in a log-log scale. Often, “diversity” refers to “channel diversity,” i.e., roughly the degree of freedom of a given channel. To attain a certain diversity order (slope) on the error probability versus SNR curve, three conditions may be satisfied: i) Transmitter 4 is well-designed so that the information symbols are encoded with sufficient redundancy (enough diversification); ii) Channel 8 is capable of providing enough degrees of freedom; iii) Receiver 4 is well designed so as to sufficiently exploit the redundancy introduced at the transmitter.

Since the diversity order G_d determines how fast the symbol error probability drops as SNR increases, G_d is to be optimized first.

We have the following theorem.

Theorem 1 (Maximum Achievable Diversity Order): For a transmitted codeword set with minimum Hamming distance δ_min, over i.i.d. FIR Rayleigh fading channels of order L, the diversity order is min(δ_min, L+1). Thus, the Maximum Achievable Diversity Order (MADO) of LE-OFDM transmissions is L+1 and in order to achieve MADO, we need δ_min≧L+1.

Proof: Since matrix A_e=V^HD_e^HD_eV in (7) is the Gram matrix¹ of D_eV, the rank r_e of A_e is the same as the rank of D_eV, which is min(δ(u, u′),L+1)≦L+1. Therefore, the diversity order of the system is

$G_d=\underset{e\in e_{}}{\min }\mathrm{rank}\left(A_c\right)=\underset{e\in e_{}}{\min }\min \left[\delta \left(u,u^{\prime }\right),L+1\right]=\min \left({\delta }_{\min },L+1\right)\le L+1,$ and the equality is achieved when δ_min≧L+1.

Theorem 1 is intuitively reasonable because the FIR Rayleigh fading channel offers us L+1 independent fading taps, which is the maximum possible number of independent replicas of the transmitted signal in the serial transmission mode. In order to achieve the MADO, any two codewords in should be different by no less than L+1 entries.

The results in Theorem 1 can also be applied to GF-coded/interleaved OFDM systems provided that channel coding or interleaving is applied only within an OFDM symbol and not across successive OFDM symbols. The diversity is again the minimum of the minimum Hamming distance of the code and L+1. To see this, it suffices to view as the codeword set of GF-coded blocks.

To achieve MADO, we need A_e to be full rank and thus positive definite for any eε_e. This is true if and only if h^HA_eh>0 for any h≠0εC^L+1. Equation (7) shows that this is equivalent to d²(y, y′)=∥D_HΘe∥²≠0, ∀eε_e, and ∀h≠0. The latter means that any two different transmitted vectors should result in different received vectors in the absence of noise, irrespective of the channel; in such cases, we call the symbols detectable or recoverable. The conditions for achieving MADO and channel-irrespective symbol detectability are summarized in the following theorem:

Theorem 2 (Symbol Detectability MADO): Under the channel conditions of Theorem 1, the maximum diversity order is achieved if and only if symbol detectability is achieved; i.e., ∥D_HΘc∥²≠0, ∀eε_e and ∀h≠0.

The result in Theorem 2 is somewhat surprising: it asserts the equivalence of a deterministic property of the code, namely symbol detectability in the absence of noise, with a statistical property, the diversity order. It can be explained though, by realizing that in random channels, the performance is mostly affected by the worst channels, despite their small realization probability. By guaranteeing detectability for any, and therefore the worst, channels, we are essentially improving the ensemble performance.

The symbol detectability condition in Theorem 2 should be checked against all pairs s and s′, which is usually not an easy task, especially when the underlying constellations are large and/or when the size K of s is large. But it is possible to identify sufficient conditions on Θ that guarantee symbol detectability and that are relatively easy to check. One such condition is provided by the following theorem.

Theorem 3 (Sufficient Condition for MADO): For i.i.d. FIR Rayleigh fading channels of order L, MADO is achieved when rank(D_HΘ)=K, ∀h≠0, which is equivalent to the following condition: Any N−L rows of Θ span the C^1×K space. The latter in turn implies that N−L≧K.

Proof: First of all, since Θ is of size N×K, it can not have rank greater than K. If MADO is not achieved, there exists at least one channel h and eε_e such that D_HΘe=0 by Theorem 2, which means that rank(D_HΘ)<K. So, MADO is achieved when D_HΘ=K. Secondly, since the diagonal entries of D_H represent frequency response of the channel h evaluated at the FFT frequencies, there can be at most L zeros on the diagonal of D_H. In order that rank(D_HΘ)=K, ∀h, it suffices to have any N−L rows of Θ span the C^1×K space. On the other hand, when there is a set of N−L rows of Θ that are linearly dependent, we can find a channel that has zeros at frequencies corresponding to the remaining L rows. Such a channel will make rank(D_HΘ)<K. This completes the proof.

The natural question that arises at this point is whether there exist LE matrices Θ that satisfy the conditions of Theorem 3. The following theorem constructively shows two classes of encoders that satisfy Theorem 3 and thus achieve MADO.

Theorem 4 (MADO-achieving encoders):

i) Vandermonde Encoders: Choose N points ρ_nεC, n=0, 1, . . . , N−1, such that ρ_m≠ρ_n, ∀n≠n. Let ρ:=[ρ₀, ρ₁, . . . , ρ_N-1]^T. Then the Vandermonde encoder Θ(ρ)εC^N×K defined by [Θ(ρ)]_n,k=ρ_n^k satisfies Theorem 3 and thus achieves MADO.

ii) Cosine Encoders: Choose N points φ₀, φ₁, . . . , φ_N-1εR, such that φ_m≠(2k+1)π and φ_m±φ_n≠2kπ, ∀m≠n, ∀kεZ. Let φ:=[φ₀, φ₁, . . . , φ_N-1]^T. Then the real cosine encoder Θ(φ)εR^N×K defined by

${\left[\Theta \left(\varphi \right)\right]}_{n,k}=\cos \left(k+\frac{1}{2}\right){\varphi }_n$ satisfies Theorem 3 and thus achieves MADO.

Proof: We first prove that Vandermonde encoders in i) satisfy the conditions of Theorem 3. Any K rows of the matrix Θ(ρ) form a square Vandermonde matrix with distinct rows. Such a Vandermonde matrix is known to have a determinant different from 0. Therefore, any K rows of Θ(ρ) are linearly independent, which satisfies the conditions in Theorem 3.

To prove Part ii) of the theorem, we show that any K rows of the encoding matrix form a non-singular square matrix. Without loss of generality, we consider the matrix formed by the first K rows:

$\begin{array}{cc}{\Theta }_1:=[\begin{array}{cccc}\cos \left(\frac{1}{2}{\varphi }_0\right)& \cos \left(\frac{3}{2}{\varphi }_0\right)& \dots & \cos \left(\frac{2K-1}{2}{\varphi }_0\right)\\ \cos \left(\frac{1}{2}{\varphi }_1\right)& \cos \left(\frac{3}{2}{\varphi }_1\right)& \dots & \cos \left(\frac{2K-1}{2}{\varphi }_1\right)\\ ⋮& ⋮& ⋮& ⋮\\ \cos \left(\frac{1}{2}{\varphi }_{K-1}\right)& \cos \left(\frac{3}{2}{\varphi }_{K-1}\right)& \dots & \cos \left(\frac{2K-1}{2}{\varphi }_{K-1}\right)\end{array}]& \left(11\right)\end{array}$

Let us evaluate the determinant det(Θ₁). Define

$z_n:=\cos \left(\frac{1}{2}{\varphi }_n\right).$ Using Chebyshev polynomials of the first kind

$T_1\left(x\right):=\cos \left(l{\cos }^{-1}x\right)=\sum _{i=0}^{\lfloor l/2\rfloor }\left(\begin{array}{c}l\\ 2i\end{array}\right){x^{1-2i}\left(x^2-1\right)}^i,$ each entry

$\cos \left(\frac{2m+1}{2}{\varphi }_n\right)$ of Θ₁ is a polynomial T_2m+1(z_n) of order 2m+1 of some

$z_n=\cos \left(\frac{1}{2}{\varphi }_n\right).$ The determinant det(Θ₁) is therefore a polynomial in z₀, . . . , z_K-1 of order Σ_n=1^K(2n−1)=K². It is easy to see that when z_n=0, or when z_m=±z_n, m≠n, Θ₁ has an all-zero row, or, two rows that are either the same or the negative of each other. Therefore, z_n, z_m−z_n, and z_m+z_n are all factors of det(Θ₁). So, g(z₀, z₁, . . . , z_K-1):=Π_nz_nΠ_m>n(z_m²−z_n²) is also a factor of det(Θ₁). But g(z₀, z₁, . . . , z_K-1) is of order K+K(K−1)=K², which means that it is different from det(Θ₁) by at most a constant. Using the leading coefficient⁴ 2^l-1 of T₁(x), we obtain the constant as Π_n=1^K2^2n-l-1=2^K(K-1); that is, det(Θ₁)=2^K(K-1)g(z₀, z₁, . . . , z_K-1).

Since φ_m≠(2k+1)π and φ_m±φ_n≠2kπ, ∀m≠n, ∀kεZ, none of z_n, z_m−z_n, and z_m+z_n can be zero. Therefore, det(Θ₁)≠0 and Θ₁ is non-singular. A similar argument can be applied to any K rows of the matrix, and the proof is complete.

Notice that up to now we have been assuming that the channel consists of i.i.d. zero-mean complex Gaussian taps. Such a model is well suited for studying average system performance in wireless fading channels, but is rather restrictive since the taps may be correlated. For correlated channels, we have the following result.

Theorem 5 (MADO of Correlated Rayleigh Channels): Let the channel h be zero-mean complex Gaussian with correlation matrix R_h. The maximum achievable diversity order equals the rank of R_h, which is achieved by any encoder that achieves MADO with i.i.d. Rayleigh channels. If R_h is full rank and MADO is achieved, then the coding advantage is different from the coding advantage in the i.i.d. case only by a constant

${\det }^{\frac{1}{L+1}}\left(R_h\right)/{\alpha }_L.$

Proof: Let r_h:=rank(R_h) and the eigen-value decomposition of R_h be

$\begin{array}{cc}{R}_h=\left[\begin{array}{cc}{U}_1& U_2\end{array}\right]\left[\begin{array}{cc}{\Lambda }_1& 0\\ 0& {\Lambda }_2\end{array}\right]\left[\begin{array}{c}{U}_1^H\\ U_2^H\end{array}\right].& \left(12\right)\end{array}$ where U₁ is (L+1)×r_h, U₂ is (L+1)×(L+1−r_h), A₁ is r_h×r_h full rank diagonal, and A₂ is an (L+1−r_h)×(L+1−r_h) all-zero matrix. Define

${\tilde{h}}_1:={\Lambda }_1^{-\frac{1}{2}}{U}_1^{H}h,{\tilde{h}}_2:=U_2^H,$ and {tilde over (h)}:=[{tilde over (h)}₁^T {tilde over (h)}₂^T]^T, where

${\Lambda }_1^{-\frac{1}{2}}$ is defined by

${\Lambda }_1^{-\frac{1}{2}}{\Lambda }_1^{-\frac{1}{2}}={\Lambda }_1^{-1}.$ Since h₂ has an autocorrelation matrix R_{{tilde over (h)}2}=U₂^HR_hU₂=A₂, all the entries of {tilde over (h)}₂ are zero almost surely. We can therefore write

$\begin{array}{cc}h=\left[\begin{array}{cc}{U}_1{\Lambda }_1^{\frac{1}{2}}& U_2\end{array}\right]\tilde{h}=U_1{\Lambda }_1^{\frac{1}{2}}{h}_1.& \left(13\right)\end{array}$ Since

$R_{{\tilde{h}}_1}={\Lambda }_1^{-\frac{1}{2}}{U}_1^H{R}_h{U}_1{\Lambda }_1^{-\frac{1}{2}}=I_{r_h}.$ the entries of {tilde over (h)}₁, which are jointly Gaussian, are i.i.d.

Substituting (13) in (7), we obtain

$\begin{array}{cc}{d}^2\left(y,y^{\prime }\right)=h^H{A}_{e}h={\tilde{h}}_1^H{\Lambda }_1^{\frac{1}{2}}{U}_1^H{A}_e{U}_1{\Lambda }_1^{\frac{1}{2}}{\tilde{h}}_1:={\tilde{h}}_1^H{\tilde{A}}_e{\tilde{h}}_1,& \left(14\right)\end{array}$ where

${\tilde{A}}_e={\Lambda }_1^{\frac{1}{2}}{U}_1^H{A}_e{U}_1{\Lambda }_1^{\frac{1}{2}}$ is an r_h×r_h matrix.

Following the same derivation as in (7)-(10), with A_e replaced by Ã_e and h replaced by {tilde over (h)}₁, we can obtain the diversity order and coding advantage for error event e as

$\begin{array}{cc}{G}_{d,e}=\mathrm{rank}\left({\tilde{A}}_e\right):={\tilde{r}}_e\le r_h\mathrm{and}{G}_{c,e}={\left(\prod _{l=0}^{{\tilde{r}}_e-1}{\tilde{\lambda }}_{e,l}\right)}^{1/{\tilde{r}}_e},& \left(15\right)\end{array}$ where {tilde over (λ)}_e,l, l=1, . . . , r_h, are the eigen-values of Ã_e.

Where Θ is designed such that MADO is achieved with i.i.d. channels, A_e is full rank for any eε_e. Then A_e is positive definite Hermitian symmetric, which means that there exists an (L+1)×(L+1) matrix B_e such that A_e=B_e^HB_e. It follows that

${\tilde{A}}_e={\Lambda }_1^{\frac{1}{2}}{U}_1^H{B}_e^H{B}_e{U}_1{\Lambda }_1^{\frac{1}{2}}$ is the Gram matrix of

$B_e{U}_1{\Lambda }_1^{\frac{1}{2}},$ and thus A_e has rank equal to

$\mathrm{rank}\left(B_e{U}_1{\Lambda }_1^{\frac{1}{2}}\right)=\mathrm{rank}\left(U_1{\Lambda }_1^{\frac{1}{2}}\right)=r_h,$ the MADO for this correlated channel.

When the MADO r_h is achieved, the coding advantage in (15) for e becomes G_c,e=det(Ã_e)^1/r_h. If in addition, R_h has full rank r_h=L+1, then det(Ã_e)^1/r_h=det(A_e)^1/(L+1)det(R_h)^1/(L+1), which means that in the full-rank correlated channel case, the full-diversity coding advantage is different from the coding advantage in the i.i.d. case only by a constant det(R_h)^1/(L+1)/α_L.

Theorem 5 asserts that the rank(R_h) is the MADO for LE-OFDM systems as well as for coded OFDM systems that do not code or interleave across OFDM symbols. Also, MADO-achieving transmissions through i.i.d. channels can achieve the MADO for correlated channels as well.

Coding advantage G_c is another parameter that needs to be optimized among the MADO-achieving encoders. Since for MADO-achieving encoders, coding advantage is given by G_c=min_e≠0G_c,e=α_L min_e≠0det(A_e), we need to maximize the minimum determinant of A_e over all possible error sequences e, among the MADO-achieving encoders.

The following theorem asserts that ZP-only transmission is one of the coding advantage maximizers.

Theorem 6 (ZP-only: maximum coding advantage): Suppose the entries of s(i) are drawn independently from a finite constellation with minimum distance of d_min(). Then the maximum coding advantage of an LE-OFDM for i.i.d. Rayleigh fading channels under as1) is G_c,max−α_Ld_min²(). The maximum coding advantage is achieved by ZP-only transmissions with any K.

In order to achieve high rate, we have adopted K=N−L and found two special classes of encoders that can achieve MADO in Theorem 4. The Vandermonde encoders are reminiscent of the parity check matrix of BCH codes, Reed-Solomon (RS) codes, and Goppa codes. It turns out that the MADO-achieving encoders and these codes are closely related.

Let us now take =C^K×1. We call the codeword set that is generated by Θ of size N×K Maximum Distance Separable (MDS) if δ_min()=N−K+1. The fact that N−K+1 is the maximum possible minimum Hamming distance of is due to the Singleton bound. Although the Singleton bound was originally proposed and mostly known for Galois field codes, its proof can be easily generalized to real/complex field as well. In our case, it asserts that δ_min≦N−K+1 when =C^K×1.

Notice that the assumption =C^K×T is usually not true in practice, because the entries of s are usually chosen from a finite-alphabet set, e.g., QPSK or QAM. But such an assumption greatly simplifies the system design task: once we can guarantee δ_min=N−K+1 for =C^K×1, we can choose any constellation from other considerations without worrying about the diversity performance. However, for a finite constellation, i.e., when δ has finite cardinality, the result on δ_min can be improved. In fact, it can be shown that even with a square and unitary K×K matrix Θ, it is possible to have δ_min=K.

To satisfy the condition in Theorem 2 with the highest rate for a given N, we need K=N−L, and δ_min=L+1=N−K+1. In other words, to achieve constellation-irrespective full-diversity with highest rate, we need the code to be MDS. According to our Theorem 4, such MDS encoders always exist for any N and K<N.

In the GF, there also exist MDS codes. Examples of GF MDS codes include single-parity-check coding, repetition coding, generalized RS coding, extended RS coding, doubly extended RS coding, algebraic-geometry codes constructed using an elliptic curve.

When a GF MDS code exists, we may use it to replace our CF linear code, and achieve the same (maximum) diversity order at the same rate. But such GF codes do not always exist for a given field and N, K. For F₂, only trivial MDS codes exist. This means that it is impossible to construct, for example, binary (and thus simply decodeable) MDS codes that have δ_min≧2, except for the repetition code. One other restriction of the GF MDS code is on the input and output alphabet. Although Reed-Solomon codes are the least restrictive among them in terms of the number of elements in the field, they are constrained on the code length and the alphabet size. Our linear encoders Θ, on the other hand, operate over the complex field with no restriction on the input symbol alphabet or the coded symbol alphabet.

We obtain analogous results on our complex field MDS codes for achieving MADO to known results for GF MDS codes.

Theorem 7 (Dual MDS code): For an MDS code generated by ΘεC^N×K, the code generated by the matrix Θ_⊥ is also MDS, where Θ_⊥ is an N×(N−K) matrix such that Θ_⊥^TΘ=0.

A generator Θ for an MDS code is called systematic if it is in the form [I_K, P]^T where P is a K×(N−K) matrix.

Theorem 8 (Systematic MDS code): A code generated by [I, P]^T is MDS if and only if every square submatrix of P is nonsingular.

To construct systematic MDS codes using Theorem 8, the following two results can be useful:

i) Every square submatrix of a Vandermonde matrix with real, positive entries is nonsingular.

ii) A K×(N−K) matrix P is called a Cauchy matrix if its (i, j)th element [P]_ij=1/(x_i+y_j) for some elements x₁, x₂, . . . , x_K, y₁, y₂, . . . , y_N-K, such that the x_i's are distinct, the y_j's are distinct, and x_i+y_j≠0 for all i, j. Any square submatrix of a Cauchy matrix is nonsingular.

Next, we discuss decoding options for our CF code. For this purpose, we restrict our attention to the case that is a finite set, e.g., a finite constellation carved from (possibly sealed and shifted) Z^K. This includes BPSK, QPSK, and QAM as special cases. Since the task of the receiver involves both channel equalization and decoding of the CF linear code, we will consider the combined task jointly and will use the words decoding, detection, and equalization interchangeably.

Maximum Likelihood Detection

To achieve MADO, LE-OFDM requires ML decoding. For the input output relationship in (2) and under the AWGN assumption, the minimum-distance detection rule becomes ML and can be formulated as follows:

$\begin{array}{cc}\hat{s}=\underset{s\in }{\arg \min }x-D_H\Theta s.& \left(16\right)\end{array}$

ML decoding of LE transmissions belongs to a general class of lattice decoding problems, as the matrix product D_HΘ in (2) gives rise to a discrete subgroup (lattice) of the C^N space under the vector addition operation. In its most general form, finding the optimum estimate in (16) requires searching over || vectors. For large block sizes and/or large constellations, it is practically impossible to perform exhaustive search since the complexity depends exponentially on the number of symbols in the block.

A relatively less complex ML search is possible with the sphere decoding (SD) algorithm (c.f., FIG. 5), which only searches coded vectors that are within a sphere centered at the received symbol x (c.f., (2)). Denote the QR decomposition of D_HΘ as D_HΘ=QR, where Q has size N×K and satisfies Q^HQ=I_K×K, and R is an upper triangular K×K matrix. The problem in (16) then converts to the following equivalent problem

$\begin{array}{cc}\hat{s}=\underset{s\in }{\arg \min }Q^{H}x-Rs,& \left(17\right)\end{array}$ SD starts its search by looking only at vectors s such that∥Q^Hx−Rs∥<C,  (18)where C is the search radius, a decoding parameter. Since R is upper triangular, in order to satisfy the inequality in (18), the last entry of s must satisfy |[R]_K,K[s]_K|<C, which reduces the search space if C is small. For one possible value of the last entry, possible candidates of the last-but-one entry are found and one candidate is taken. The process continues until a vector of s₀ is found that satisfies (18). Then the search radius C is set equal to ∥Q^Hx−Rs₀∥ and a new search round is started. If no other vector is found inside the radius, then s₀ is the ML solution. Otherwise, if s₁ is found inside the sphere, the search radius is again reduced to ∥Q^Hx−Rs₁∥, and so on. If no s₀ is ever found inside the initial sphere of radius C, then C is too small. In this case, either a decoding failure is declared or C is increased.

The complexity of the SD is polynomial in K, which is better than exponential but still too high for practical purposes. Indeed, it is not suitable for codes of block size greater than, say, 16. When the block size is small, the sphere decoder can be considered as an option to achieve the ML performance at manageable complexity.

In the special case of ZP-only transmissions, the received vector is given by {tilde over (x)}=Hs+ñ. Thanks to the zero-padding, the full convolution of the transmitted block s with the FIR channel is preserved and the channel is represented as the banded Toeplitz matrix H. In such a case, Viterbi decoding can be used at a complexity of (Q^L) per symbol, where Q is the constellation size of the symbols in s.

Low-Complexity Linear Detection

Zero-forcing (ZF) and MMSE detectors (equalizers) offer low-complexity alternatives. The ZF and MMSE equalizers based on the input-output relationship (2) can be written as:G^zf=(D_HΘ)^T and G^mmse=R_sΘ^HD_H^H(σ_η²I_N+D_HΘR_sΘ^HD_H^H)⁻¹,respectively, where (·)^T denotes pseudo-inverse, σ_η² is the variance of entries of noise η, and R_s is the autocorrelation matrix of s. Given the ZF and MMSE equalizers, they each require (N×K) operations per K symbols. So per symbol, they require only (N) operations. To obtain the ZF or MMSE equalizers, inversion of a N×N matrix is involved, which has complexity (N³). However, the equalizers only needs to be recomputed when the channel changes.

Decision-Directed Detection

The ML detection schemes in general have high complexity, while the linear detectors may have decreased performance. The class of decision-directed detectors lies between these categories, both in terms of complexity and in terms of performance.

Decision-directed detectors capitalize on the finite alphabet property that is almost always available in practice. In the equalization scenario, they are more commonly known as Decision Feedback Equalizers (DFE). In a single-user block formulation, the DFE has a structure as shown in FIG. 6, where the feed-forward filter is represented as a matrix W and the feedback filter is presented as B. Since we can only feed back decisions in a casual fashion, B is usually chosen to be a strictly upper or lower triangular matrix with zero diagonal entries. Although the feedback loop is represented as a matrix, the operations happen in a serial fashion: the estimated symbols are fed back serially as their decisions are formed one by one. The matrices W and B can be designed according to ZF or MMSE criteria. When B is chonse to be triangular and the MSE between the block estimate before the decision device is minimized, the feed-forward and feedback filtering matrices can be found from the following equations:R_s⁻¹+Θ^HD_H^HR_η⁻¹D_HΘ=U^HΛU,  (19)W=UR_sΘ^HD_H^H(R_η+D_HΘR_sΘ^HD_H^H)⁻¹, B=U−I,  (20)where the R's denote autocorrelation matrices, (19) was obtained using Cholesky decomposition, and U is an upper triangular matrix with unit diagonal entries. Since the feed-forward and feedback filtering entails only matrix-vector multiplications, the complexity of such decision directed schemes is comparable to that of linear detectors. Because decision directed schemes capitalize on the finite-alphabet property of the information symbols, the performance is usually (much) better than linear detectors.

As an example, we list in the following table the approximate number of flops needed for different decoding schemes when K=14, L=2, N=16, and BPSK modulation is deployed; i.e., ={±1}^K.

TABLE 1
Decoding Scheme     order of Flops/symbol
Exhaustive ML       >2K = 214 = 16.384
Sphere Decoding     ≈800 (empirical)
ZF/MMSE             ≈N = 16
Decision-Directed   ≈N = 16
Viterbi for ZP-only 2L = 22 = 4

Iterative Detectors

Other possible decoding methods include iterative detectors, such as successive interference cancellation with iterative least squares (SIC-ILS), and multistage cancellations. These methods are similar to the illustrated DFE in the interference from symbols that are decided in a block is canceled before a decision on the current symbol is made. In SIC-ILS, least squares is used as the optimization criterion and at each step or iteration, the cost function (least-squares) will decrease or remain the same. In multistage cancellation, the MMSE criterion is often used such that MF is optimum after the interference is removed (supposing that the noise is white). The difference between a multistage cancellation scheme and the block DFE is that the DFE symbol decisions are made serially; and for each undecided symbol, only interference from symbols that have been decided is cancelled; while in multistage cancellation, all symbols are decided simultaneously and then their mutual interferences are removed in a parallel fashion.

As illustrated in FIG. 7, another embodiment may utilize for LEOFDM equalization an iterative “sum-product” decoding algorithm, which is also used in Turbo decoding. In particular, the coded system is represented using a factor graph, which describes the interdependence of the encoder input, the encode output, and the noise-corrupted coded symbols.

As a simple example, suppose the encoder takes a block of 3 symbols s:=[s₀, s₁, s₂]^T as input and linearly encodes them by a 4×3 matrix Θ to produce the coded symbols u:=[u₀, u₁, u₂, u₃]. After passing through the channel (OFDM modulation/demodulation), we obtain the channel output x_i=H(ε^j2πi/4)u_i, i=0, 1, 2, 3. The factor graph for such a coded system is shown in FIG. 7, where the LE is represented by linear constraints between the LE input symbols s and the LE output symbols u.

Parallel Encoding for Low Complexity Decoding

When the number of carriers N is very large (e.g., 1,024), it is desirable to keep the decoding complexity manageable. To achieve this we can split the encoder into several smaller encoders. Specifically, we can choose Θ=PΘ′, where P is a permutation matrix that interleaves the subcarriers, and Θ′ is a block diagonal matrix: Θ′=diag(Θ₀, Θ₁, . . . , Θ_M-1). This is a essentially a form of coding for interleaved OFDM, except that the coding is done in complex domain here. The matrices Θ_m, m=0, . . . , M−1 are of smaller size than Θ and all of them can even be chosen to be identical. With such designed Θ, decoding s from the noisy D_HΘs is equivalent to decoding M coded sub-vectors of smaller sizes and therefore the overall decoding complexity can be reduced considerably. Such a decomposition is particularly important when a high complexity decoder such as the sphere decoder is to be deployed.

The price paid for low decoding complexity is a decrease in transmission rate. When such parallel encoding is used, we should make sure that each of the Θ_m matrices can guarantee full diversity, which requires Θ_m to have L redundant rows. The overall Θ will then have ML redundant rows, which corresponds to an M-fold increase of the redundancy of a full single encoder of size N×K. If a fixed constellation is used for entries in s, then square Θ_m's can be used, which does not lead to loss of efficiency.

FIGS. 8-10 are graphs that illustrate exemplary results of simulations of the described techniques. In the illustrated results, we compare the proposed wireless communication techniques with existing coded OFDM systems that deploy existing GF block codes and convolutional codes. In all cases, BPSK constellation is used, and in Test Case 2 and 3, the binary encoded symbols are mapped to ±1's before OFDM modulation.

Test case 1 (Decoding of LE-OFDM): We first test the performance of different decoding algorithms. The LE-OFDM system has parameters K=14, N=16, L=2. The channel is i.i.d. Rayleigh and BER's for 200 random channel realizations according to As1) are averaged. FIG. 8 shows the performance of ZF, MMSE, DFE, and sphere decoding (ML) for LE-OFDM. We notice that at BER of 10⁻⁴ DFE performs about 2 dB better than the MMSE detectors, while at the same time it is only less than 1 dB inferior to the sphere decoder, which virtually achieves the ML decoding performance. The complexity of ZF, MMSE, DFE is all about N=16 flops per symbol, which is much less than the sphere decoding algorithm, which empirically needs about 800 flops per symbol in this case.Test case 2 (Comparing LE-OFDM with BCH-coded OFDM): For demonstration and verification purposes, we first compare LE-OFDM with coded OFDM that relies on GF block coding. The channel is modeled as FIR with 5 i.i.d. Rayleigh distributed taps. In FIG. 9, we illustrate Bit Error Rate (BER) performance of CF coded OFDM with Vandermonde code of Theorem 4, and that of binary BCH-coded OFDM. The system parameters are K=26, N=31. The generating polynomial of the BCH code is g(D)=1+D²+D⁵. Since we can view this BCH as a rate 1 convolutional code with the same generator and with termination after 26 information symbols (i.e., the code ends at the all-zero state), we can use the Viterbi algorithm for soft-decision ML BCH decoding. For LE-OFDM, since the transmission is essentially a ZP-only single-carrier scheme, the Viterbi algorithm is also applicable for ML decoding.

Since the binary (26, 31) BCH code has minimum Hamming distance 3, it possesses a diversity order of 3, which is only half of the maximum possible (L+1=6) that LE-OFDM achieves with the same spectral efficiency. This explains the difference in their performance. We can see that when the optimum ML decoder is adopted by both receivers. LE-OFDM outperforms coded OFDM with BCH coding considerably. The slopes of the corresponding BER curves also confirm our theoretical results.

Test case 3 (Comparing LE-OFDM with convolutionally coded OFDM): In this test, we compare (see FIG. 10) our LE-OFDM system with convolutionally coded OFDM (with a rate ½ code punctured to rate ¾ followed by interleaving) that is deployed by the HiperLAN2 standard over the channels used in Test Case 2. The rate ½ mother code has its generator in octal form as (133, 171) and there are 64 states in its trellis. Every 3rd bit from the first branch and every second bit from the second branch of the mother code are punctured to obtain the rate ¾ code, which results in a code whose weight enumerating function is 8W⁵+31W⁶+160W⁷+ . . . . So the free distance is 5, which means that the achieved diversity is 5, less than the diversity order 6 achieved by LE-OFDM.

The parameters are K=36, N=48. We use two parallel truncated DCT encoders; that is, Θ=I_2×2Θ₀, where denotes Kronecker product, and Θ₀ is a 24×18 encoder obtained by taking the first 18 columns of a 24×24 DCT matrix. With ML decoding, LE-OFDM performs about 2 dB better than convolutionally coded OFDM. From the ML performance curves in FIG. 10, LE-OFDM seems to achieve a larger coding advantage than the punctured convolutional code we used.

Surprisingly, even with linear MMSE equalization, the performance of LE-OFDM is better than coded OFDM for SNR values less than 12 dB. The complexity of ML decoding for LE-OFDM is quite high in the order of 1,000 flops per symbol. But the ZF and MMSE decoders have comparable or even lower complexity than the Viterbi decoder for the convolutional code.

The complexity of LE-OFDM can be dramatically reduced using the parallel encoding method with square encoders. It is also possible to combine CF coding with conventional GF coding, in which case only small square encoders of size 2×2 or 4×4 are necessary to achieve near optimum performance.

Various embodiments of the invention have been described. The described techniques can be embodied in a variety of receivers and transmitters including base stations, cell phones, laptop computers, handheld computing devices, personal digital assistants (PDA's), and the like. The devices may include a digital signal processor (DSP), field programmable gate array (FPGA), application specific integrated circuit (ASIC) or similar hardware, firmware and/or software for implementing the techniques. If implemented in software, a computer readable medium may store computer readable instructions, i.e., program code, that can be executed by a processor or DSP to carry out one of more of the techniques described above. For example, the computer readable medium may comprise random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), electrically erasable programmable read-only memory (EEPROM), flash memory, or the like. The computer readable medium may comprise computer readable instructions that when executed in a wireless communication device, cause the wireless communication device to carry out one or more of the techniques described herein. These and other embodiments are within the scope of the following claims.

Claims

1. A wireless communication device comprising: an encoder that applies a linear transformation to a stream of information bearing symbols selected from a constellation having a finite alphabet to produce a stream of precoded symbols that are complex numbers and are not restricted by the constellation of the information bearing symbols; anda modulator to produce an output waveform in accordance with the stream of precoded symbols for transmission through a wireless channel.

2. The wireless communication device of claim 1, wherein the modulator generates the output waveform as a multicarrier waveform having a set of subcarriers, and the encoder encodes the stream of information bearing symbols so that the subcarriers carry different linear combinations of the information symbols.

3. The wireless communication device of claim 1, wherein the encoder applies the linear transformation by applying a unitary matrix to the information bearing symbols.

4. A wireless communication device comprising: an encoder that applies a matrix to linearly transform blocks of K information bearing symbols selected from a constellation having a finite alphabet to produce blocks of N precoded symbols that are complex numbers and are not restricted to the constellation of the information bearing symbols; anda modulator that generates a multicarrier waveform having a set of subcarriers, where N is the number of subcarriers of the multi-carrier waveform and K is less than or equal to N.

5. The wireless communication device of claim 4, wherein the linear encoder has a code rate r=K/N.

6. The wireless communication device of claim 4, wherein the linear encoder applies a matrix of size N×K to blocks of K information bearing symbols to produce blocks of N precoded symbols.

7. A wireless communication device, comprising: an encoder that applies a matrix to linearly transform blocks of K information bearing symbols selected from a constellation having a finite alphabet to produce blocks of N precoded symbols that are complex numbers and are not restricted by the constellation of the information bearing symbols; anda modulator that generates a multicarrier waveform having a set of subcarriers for transmission over a wireless channel,wherein N is the number of subcarriers and K is less than or equal to N, andwherein the size of the matrix is selected as a function of an order L of the wireless channel, and the number K of symbols per block is selected as a function of the channel order L.

8. The wireless communication device of claim 7, wherein K is selected so that K≦N−L.

9. The wireless communication device of claim 7, wherein K is selected so that K=N−L.

10. The wireless communication device of claim 6, wherein the linear encoder applies the matrix to perform a vector multiplication on the blocks of K information bearing symbols to produce blocks of N precoded symbols, and applies each block of N precoded symbols across the N subcarriers.

11. The wireless communication device of claim 1, wherein the wireless communication device comprises one of a base station and a mobile device.

12. A wireless communication device comprising: an encoder that applies a plurality of M matrices to linearly transform a stream of information bearing symbols selected from a constellation having a finite alphabet to produce a stream of precoded symbols that are complex numbers and are not restricted by the constellation of the information bearing symbols; anda modulator to produce an output waveform in accordance with the stream of precoded symbols for transmission through a wireless channel,where the matrices are identical and collectively have M*L redundant rows, where L represents an order of the channel.

13. A wireless communication device comprising: a demodulator that receives a waveform carrying a encoded transmission and produces a demodulated data stream, wherein the encoded data stream was produced by applying a linear transformation to a stream of information bearing symbols selected from a constellation having a finite alphabet to produce a stream of precoded symbols that are complex numbers and are not restricted by the constellation of the information bearing symbols; anda decoder that decodes the demodulated data to produce estimated data.

14. The wireless communication device of claim 13, wherein the decoder applies one of maximum-likelihood detection, zero-force (ZF) detection, minimum mean squared error (MMSE) detection, decision-directed detection, iterative detection, to decode the demodulated data.

15. The wireless communication device of claim 13, wherein the wireless communication device comprises one of a base station and a mobile device.

16. A method comprising: applying a linear transformation to a stream of information bearing symbols selected from a constellation having a finite alphabet to produce a stream of precoded symbols that are complex numbers and are not restricted by the constellation of the information bearing symbols; andoutputting a waveform in accordance with the stream of precoded symbols for transmission through a wireless channel.

17. The method of claim 16, wherein outputting the waveform comprises: outputting the output waveform as a multicarrier waveform having a set of subcarriers; andencoding the stream of information bearing symbols so that the subcarriers carry different linear combinations of information symbols.

18. The method of claim 16, wherein applying the linear transformation to the stream of information bearing symbols comprises applying a unitary matrix to the stream of information bearing symbols.

19. A method comprising: applying a matrix to linearly transform blocks of K information bearing symbols that are selected from a constellation having a finite alphabet to produce blocks of N precoded symbols that are complex numbers and are not restricted by the constellation of the information bearing symbols, andoutputting a multicarrier waveform having a set of subcarriers in accordance with the stream of precoded symbols for transmission through a wireless channel,where N is the number of subcarriers, and K is less than or equal to N.

20. The method of claim 19, wherein applying the linear transformation comprises applying the linear transformation to blocks of K information bearing symbols to produce blocks of N precoded symbols at a code rate r=K/N.

21. The method of claim 19, wherein applying the linear transformation comprises applying a matrix of size N×K to the blocks of information bearing symbols.

22. The method of claim 19, further comprising selecting the number of symbols per block K as a function of an order of the channel.

23. The method of claim 19, further comprising selecting the number of symbols per block so that K≦N−L, wherein L represents an order of the channel.

24. The method of claim 19, further comprising selecting the number of symbols per block so that K=N−L, where L represents an order of the channel.

25. The method of claim 19, wherein applying the linear transformation comprises applying a matrix to perform a vector multiplication on the blocks of K information bearing symbols to produce blocks of N precoded symbols.

26. A method comprising: applying a plurality of M matrices to linearly transform a stream of information bearing symbols selected from a constellation having a finite alphabet, wherein the M matrices linearly transform the stream of information bearing symbols to produce a stream of precoded symbols that are complex numbers and are not restricted by the constellation of the information bearing symbols; andoutputting a waveform in accordance with the stream of precoded symbols for transmission through a wireless channel,where the matrices are identical and have M*L redundant rows and, where L represents an order of the channel.

27. A computer-readable medium comprising instructions to cause a programmable processor to: apply a linear transformation to a stream of information bearing symbols selected from a constellation having a finite alphabet to produce a stream of precoded symbols that are complex numbers and are not restricted by the constellation of the information bearing symbols; andoutput waveform in accordance with the stream of precoded symbols for transmission through a wireless channel.

28. The computer-readable medium of claim 27, further comprising instructions to cause the programmable processor to: output the output waveform as a multicarrier waveform having a set of subcarriers; andencode the stream of information bearing symbols so that the subcarriers carry different linear combinations of information symbols.