Continuous beamforming for a MIMO-OFDM system

Abstract

A transmitting entity performs spatial processing on data symbols for each subband with an eigenmode matrix, a steering matrix, or an identity matrix to obtain spatially processed symbols for the subband. The data symbols may be sent on orthogonal spatial channels with the eigenmode matrix, on different spatial channels with the steering matrix, or from different transmit antennas with the identity matrix. The transmitting entity further performs beamforming on the spatially processed symbols, in the frequency domain or time domain, prior to transmission from the multiple transmit antennas. A receiving entity performs the complementary processing to recover the data symbols sent by the transmitting entity. The receiving entity may derive a spatial filter matrix for each subband based on a MIMO channel response matrix for that subband and perform receiver spatial processing for the subband with the spatial filter matrix.

Description

BACKGROUND

1. Field

The present invention relates generally to communication, and more specifically to data transmission in a multiple-input multiple-output (MIMO) communication system that utilizes orthogonal frequency division multiplexing (OFDM).

2. Background

A MIMO system employs multiple (T) transmit antennas at a transmitting entity and multiple (R) receive antennas at a receiving entity for data transmission. A MIMO channel formed by the T transmit antennas and R receive antennas may be decomposed into S spatial channels, where S≦min {T, R}. The S spatial channels may be used to transmit data in parallel to achieve higher throughput and/or redundantly to achieve greater reliability.

OFDM is a multi-carrier modulation technique that effectively partitions the overall system bandwidth into multiple (K) orthogonal frequency subbands. These subbands are also referred to as tones, subcarriers, bins, and frequency channels. With OFDM, each subband is associated with a respective subcarrier that may be modulated with data.

A MIMO-OFDM system is a MIMO system that utilizes OFDM. The MIMO-OFDM system has S spatial channels for each of the K subbands. Each spatial channel of each subband may be called a “transmission channel”. Each transmission channel may experience various deleterious channel conditions such as, e.g., fading, multipath, and interference effects. The transmission channels for the MIMO channel may also experience different channel conditions and may achieve different signal-to-noise-and-interference ratios (SNRs). The SNR of each transmission channel determines its transmission capacity, which is typically quantified by a particular data rate that may be reliably transmitted on the transmission channel. For a time variant wireless channel, the channel conditions change over time and the SNR of each transmission channel also changes over time. The different SNRs for different transmission channels plus the time varying nature of the SNR for each transmission channel make it challenging to efficiently transmit data in a MIMO system.

If the transmitting entity has knowledge of the channel condition, then it may transmit data in a manner to more fully utilize the capacity of each transmission channel. However, if the transmitting entity does not know the channel condition, then it may need to transmit data at a low rate so that the data transmission can be reliably decoded by the receiving entity even with the worst-case channel condition. Performance would then be dictated by the expected worst-case channel condition, which is highly undesirable.

There is therefore a need in the art for techniques to more efficiently transmit data in a MIMO-OFDM system, especially when the channel condition is not known by the transmitting entity.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a MIMO-OFDM system with an access point and user terminals.

FIG. 2 shows a block diagram of a transmitting entity and a receiving entity.

FIG. 3 shows an OFDM waveform in the frequency domain.

FIGS. 4 and 5 show a transmit (TX) spatial processor with a frequency-domain beamformer.

FIG. 6 shows a block diagram of an OFDM modulator.

FIG. 7 shows a TX spatial processor with a time-domain beamformer.

FIG. 8A shows a time-domain beamformer with circular shifting.

FIG. 8B shows transmissions with the beamformer in FIG. 8A.

FIG. 9A shows a time-domain beamformer with linear delay.

FIG. 9B shows transmissions with the beamformer in FIG. 9A.

FIG. 10 shows plots of linear phase shifts across subbands for four antennas.

DETAILED DESCRIPTION

The word “exemplary” is used herein to mean “serving as an example, instance, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

FIG. 1 shows a MIMO-OFDM system 100 with an access point (AP) 110 and user terminals (UTs) 120. An access point is generally a fixed station that communicates with the user terminals and may also be referred to as a base station or some other terminology. A user terminal may be fixed or mobile and may also be referred to as a mobile station, a wireless device, a user equipment (UE), or some other terminology. For a centralized network architecture, a system controller 130 couples to the access points and provides coordination and control for these access points.

Access point 110 is equipped with multiple antennas for data transmission and reception. Each user terminal 120 is also equipped with multiple antennas for data transmission and reception. A user terminal may communicate with the access point, in which case the roles of access point and user terminal are established. A user terminal may also communicate peer-to-peer with another user terminal.

FIG. 2 shows a block diagram of a transmitting entity 210 and a receiving entity 250 in system 100. Transmitting entity 210 is equipped with multiple (T) transmit antennas and may be an access point or a user terminal. Receiving entity 250 is equipped with multiple (R) antennas and may also be an access point or a user terminal.

At transmitting entity 210, a TX data processor 212 processes (e.g., encodes, interleaves, and symbol maps) traffic/packet data to generate data symbols. As used herein, a “data symbol” is a modulation symbol for data, a “pilot symbol” is a modulation symbol for pilot (which is data that is known a priori by both the transmitting and receiving entities), a “transmit symbol” is a symbol to be sent on one subband of one transmit antenna, and a “received symbol” is a symbol obtained on one subband of one receive antenna. A TX spatial processor 220 receives and demultiplexes pilot and data symbols onto the proper subbands, performs spatial processing as described below, and provides T streams of transmit symbols for the T transmit antennas. A modulator (MOD) 230 performs OFDM modulation on each of the T transmit symbol streams and provides T streams of time-domain samples to T transmitter units (TMTR) 232a through 232t. Each transmitter unit 232 processes (e.g., converts to analog, amplifies, filters, and frequency upconverts) its sample stream to generate a modulated signal. Transmitter units 232a through 232t provide T modulated signals for transmission from T antennas 234a through 234t, respectively.

At receiving entity 250, R antennas 252a through 252r receive the T transmitted signals, and each antenna 252 provides a received signal to a respective receiver unit (RCVR) 254. Each receiver unit 254 processes its received signal and provides a stream of input samples to a corresponding demodulator (DEMOD) 260. Each demodulator 260 performs OFDM demodulation on its input sample stream to obtain receive data and pilot symbols, provides the received data symbols to a receive (RX) spatial processor 270, and provides the received pilot symbols to a channel estimator 284 within a controller 280. Channel estimator 284 derives a channel response estimate for an actual or effective MIMO channel between transmitting entity 210 and receiving entity 250 for each subband used for data transmission. Controller 280 derives spatial filter matrices based on the MIMO channel response estimates. RX spatial processor 270 performs receiver spatial processing (or spatial matched filtering) on the received data symbols for each subband with the spatial filter matrix derived for that subband and provides detected data symbols for the subband. The detected data symbols are estimates of the data symbols sent by transmitting entity 210. An RX data processor 272 then processes the detected data symbols for all subbands and provides decoded data.

Controllers 240 and 280 direct the operation of the processing units at transmitting entity 210 and receiving entity 250, respectively. Memory units 242 and 282 store data and/or program code used by controllers 240 and 280, respectively.

FIG. 3 shows an OFDM waveform in the frequency domain. OFDM provides K total subbands, and the subcarrier for each subband may be individually modulated with data. Of the K total subbands, N_D subbands may be used for data transmission, N_P subbands may be used for pilot transmission, and the remaining N_G subbands may be unused and serve as guard subbands, where K=N_D+N_P+N_G. In general, system 100 may utilize any OFDM structure with any number of data, pilot, guard, and total subbands. For simplicity, the following description assumes that all K subbands are usable for data and pilot transmission.

System 100 may support data transmission using multiple operating modes. Each operating mode utilizes different spatial processing at the transmitting entity. In an embodiment, each operating mode may utilize (1) “eigensteering” to transmit data symbols on orthogonal spatial channels (or “eigenmodes”) of a MIMO channel, (2) “matrix steering” to transmit each data symbol on all S spatial channels of the MIMO channel, or (3) no spatial processing to transmit each data symbol from one transmit antenna. Eigensteering is also called eigenmode transmission or full channel state information (full-CSI) transmission. Matrix steering may be used to achieve spatial diversity. Data transmission without spatial processing is also called partial-CSI transmission. In an embodiment, each operating mode may or may not utilize beamforming to introduce additional diversity for the T sample streams sent from the T transmit antennas.

The operating mode with the combination of matrix steering and beamforming is called “spatial spreading”. The transmitting entity may use spatial spreading to achieve spatial and frequency/time diversity, for example, if the transmitting entity does not know the MIMO channel response.

Transmitter Spatial Processing

In system 100, the MIMO channel formed by the T transmit antennas at transmitting entity 210 and the R receive antennas at receiving entity 250 may be characterized by an R×T channel response matrix H(k) for each subband k, which may be given as:

$\begin{array}{cc}\underset{\_}{H}\left(k\right)=\left[\begin{array}{cccc}{h}_{0,0}\left(k\right)& h_{0,1}\left(k\right)& \dots & h_{0,T-1}\left(k\right)\\ h_{1,0}\left(k\right)& h_{1,1}\left(k\right)& \dots & h_{1,T-1}\left(k\right)\\ ⋮& ⋮& \ddots & ⋮\\ h_{R-1,0}\left(k\right)& h_{R-1,1}\left(k\right)& \dots & h_{R-1,T-1}\left(k\right)\end{array}\right],\mathrm{for}k=0,\dots ,K-1,& \mathrm{Eq}.\left(1\right)\end{array}$ where entry h_i,j(k), for i=0, . . . , R−1 and j=0, . . . , T−1, denotes the coupling or complex channel gain between transmit antenna j and receive antenna i for subband k. For simplicity, the MIMO channel is assumed to be full rank with S=T≦R.

For data transmission with eigensteering, eigenvalue decomposition may be performed on a correlation matrix of H(k) to obtain S eigenmodes of H(k), as follows:R(k)=H^H(k)·H(k)=E(k)·Λ(k)·E^H(k)  Eq. (2)where

R(k) is a T×T correlation matrix of H(k);

E(k) is a T×T unitary matrix whose columns are eigenvectors of R(k);

Λ(k) is a T×T diagonal matrix of eigenvalues of R(k); and

“^H” denotes a conjugate transpose.

A unitary matrix U is characterized by the property U^H·U=I, where I is the identity matrix. The columns of a unitary matrix are orthogonal to one another, and each column has unit power. The matrix E(k) is also called an “eigenmode” matrix or a “transmit” matrix and may be used for spatial processing by the transmitting entity to transmit data on the S eigenmodes of H(k). The eigenmodes may be viewed as orthogonal spatial channels obtained through decomposition. The diagonal entries of Λ(k) are eigenvalues of R(k), which represent the power gains for the S eigenmodes. The eigenvalues in Λ(k) may be ordered from largest to smallest, and the columns of E(k) may be ordered correspondingly. Singular value decomposition may also be performed to obtain matrices of left and right eigenvectors, which may be used for eigensteering.

For data transmission with eigensteering, the transmitting entity may perform spatial processing for each subband k as follows:z_es(k)=E(k)·s(k),  Eq. (3)where s(k) is a vector with up to S data symbols to be sent on subband k; and

z _es(k) is a vector with T spatially processed symbols for subband k.

In general, D data symbols may be sent simultaneously on D(best) eigenmodes of H(k) for each subband k, where 1≦D≦S. The D data symbols in s(k) are spatially processed with D columns of E(k) corresponding to the D selected eigenmodes.

For data transmission with matrix steering, the transmitting entity may perform spatial processing for each subband k as follows:z_ss(k)=V(k)·s(k),  Eq. (4)where V(k) is a unitary steering matrix for subband k; and

z _ss(k) is a vector with up to T spread symbols for subband k.

Each data symbol in s(k) is multiplied with a respective column of V(k) to obtain up to T spread symbols. The steering matrix V(k) may be generated in a manner to simplify the matrix multiplication in equation (4), as described below.

In general, D data symbols may be sent simultaneously on each subband k with matrix steering, where 1≦D≦S. The D data symbols in s(k) may be multiplied with a T×D unitary steering matrix V(k) to obtain T spatially processed symbols for z_ss(k). Each spatially processed symbol for each subband k includes a component of each of the D data symbols being sent on subband k. The T spatially processed symbols for each subband k are then transmitted on the S spatial channels of H(k).

For partial-CSI transmission, the transmitting entity may perform spatial processing for each subband k as follows:z_pcsi(k)=s(k),  Eq. (5)where z_pcsi(k) is a vector with up to T data symbols to be sent on subband k. In effect, the transmitting entity performs spatial processing with the identity matrix I for partial-CSI transmission.

The transmitting entity thus spatially processes the data vector s(k) for each subband k to obtain a corresponding vector z(k) of spatially processed symbols for that subband. The vector z(k) is equal to z_es(k) for eigensteering, z_ss(k) for matrix steering, and z_pcsi(k) for partial-CSI transmission.

Beamforming

The transmitting entity may selectively perform beamforming on the vector z(k) for each subband k, as follows:x(k)=B(k)·z(k),  Eq. (6)where B(k) is a T×T beamforming matrix for subband k; and

x(k) is a vector with T transmit symbols to be sent from the T transmit antennas for subband k.

If beamforming is not performed, then the beamforming matrix B(k) is replaced with the identity matrix I in equation (6).

The transmit vector x_bes(k) for eigensteering with beamforming may be expressed as:x_bes(k)=B(k)·E(k)·s(k).  Eq. (7)

The transmit vector x_bss(k) for spatial spreading, which is matrix steering with beamforming, may be expressed as:x_bss(k)=B(k)·V(k)·s(k).  Eq. (8)A matrix B(k)·V(k) may be pre-computed for each subband k. In this case, the transmit vector x_bss(k) may be obtained with a single matrix multiply. The matrices V(k) and B(k) may also be applied in two steps and possibly in different manners. For example, the matrix V(k) may be applied in the frequency domain with a matrix multiply and the matrix B(k) may be applied in the time domain with circular or linear delays, as described below.

The transmit vector x_bns(k) for partial-CSI transmission with beamforming may be expressed as:x_bns(k)=B(k)·s(k).  Eq. (9)

The beamforming matrix B(k) for each subband k is a diagonal matrix having the following form:

$\begin{array}{cc}\underset{\_}{B}\left(k\right)=\left[\begin{array}{cccc}{b}_0\left(k\right)& 0& \dots & 0\\ 0& b_1\left(k\right)& \dots & 0\\ ⋮& ⋮& \ddots & ⋮\\ 0& 0& \dots & b_{T-1}\left(k\right)\end{array}\right],\mathrm{for}k=0,\dots ,K-1,& \mathrm{Eq}.\left(10\right)\end{array}$ where b_i(k) is a weight for subband k of transmit antenna i. As shown in equation (6), the i-th element of z(k) is multiplied by the i-th diagonal weight in B(k).

The beamforming matrices for the K subbands may be defined such that continuous beamforming is achieved across the K subbands. The beamforming matrix B(k) for each subband k defines an antenna beam for that subband. K different beamforming matrices may be used for the K subbands to obtain different antenna beams across the subbands. The K beamforming matrices may be varied in a continuous manner (instead of an abrupt or discontinuous manner) so that the antenna beams change in a continuous manner across the K subbands. Continuous beamforming thus refers to a continuous change in the antenna beams across the K subbands.

In an embodiment, the weights in the beamforming matrix B(k) for each subband k are defined as follows:

$\begin{array}{cc}{b}_i\left(k\right)=g\left(i\right)·e^{j\frac{2\pi ·i·k}{K}},\mathrm{for}i=0,\dots ,T-1\mathrm{and}k=0,\dots ,K-1,& \mathrm{Eq}.\left(11\right)\end{array}$ where g(i) is a complex gain for transmit antenna i.

The magnitude of the complex gain for each transmit antenna may be set to one, or ∥g(i)∥=1.0 for i=0, . . . , T−1. The weights shown in equation (11) correspond to a progressive phase shift across the K subbands of each transmit antenna, with the phase shift changing at different rates for the T transmit antennas. These weights effectively form a different beam for each subband for a linear array of T equally spaced antennas.

In a specific embodiment, the weights are defined as follows:

$\begin{array}{cc}{b}_i\left(k\right)=e^{-\mathrm{j\pi }·i}·e^{j^{\frac{2\pi ·i·k}{K}}}=e^{j2\pi \frac{i}{K}·\left(k-\frac{K}{2}\right)},& \mathrm{Eq}.\left(12\right)\end{array}$ for i=0, . . . , T−1 and k=0, . . . , K−1. The embodiment shown in equation (12) uses g(i)=e^−jπ·i for equation (11). This results in a phase shift of zero being applied to subband K/2+1 for each antenna.

FIG. 10 shows plots of the phase shifts for each transmit antenna for a case with T=4. The center of the K subbands is typically considered to be at zero frequency. The weights generated based on equation (12) may be interpreted as creating a linear phase shift across the K subbands. Each transmit antenna i, for i=0, . . . , T−1, is associated with a phase slope of 2π·i/K. The phase shift for each subband k, for k=0, . . . , K−1, of transmit antenna i is given as 2π·i·(k−K/2)/K. The use of g(i)=e^−jπ·i result in subband k=K/2 observing a phase shift of zero.

The weights derived based on equation (12) may be viewed as a linear filter having a discrete frequency response of G_i(k′) for each transmit antenna i. This discrete frequency response may be expressed as:

$\begin{array}{cc}{G}_i\left(k^{\prime }\right)=b_i\left(k^{\prime }+K/2\right)=e^{j2\pi \frac{i·k^{\prime }}{K}},& \mathrm{Eq}.\left(13\right)\end{array}$ for i=0, . . . , T−1 and k′=(−K/2) . . . (K/2−1). Subband index k is for a subband numbering scheme that places the zero frequency at subband N_center=K/2. Subband index k′ is a shifted version of subband index k by K/2, or k′=k−K/2. This results in subband zero being at zero frequency for the new subband numbering scheme with index k′. N_center may be equal to some other value instead of K/2 if index k is defined in some other manner (e.g., k=1, . . . , K) or if K is an odd integer value.

A discrete time-domain impulse response g_i(n) for the linear filter may be obtained by performing a K-point inverse discrete Fourier transform (IDFT) on the discrete frequency response G_i(k′). The impulse response g_i(n) may be expressed as:

$\begin{array}{cc}\begin{array}{c}{g}_i\left(n\right)=\frac{1}{K}·\sum _{k^{\prime }=-K/2}^{K/2-1}{G}_i\left(k^{\prime }\right)·e^{j2\pi \frac{n·k^{\prime }}{K}},\\ =\frac{1}{K}·\sum _{k^{\prime }=-K/2}^{K/2-1}{e}^{j2\pi \frac{i·k^{\prime }}{K}}·e^{j2\pi \frac{n·k^{\prime }}{K}},\\ =\frac{1}{K}·\sum _{k^{\prime }=-K/2}^{K/2-1}{e}^{j2\pi \frac{k^{\prime }}{K}\left(i+n\right)},\\ =\{\begin{array}{cc}1& \mathrm{for}n=-i\\ 0& \mathrm{otherwise}\end{array}\end{array}& \mathrm{Eq}.\left(14\right)\end{array}$ where n is an index for sample period and has a range of n=0, . . . , K−1. Equation (14) indicates that the impulse response g_i(n) for transmit antenna i has a single tap with unit-magnitude at a delay of i sample periods and is zero at all other delays.

Beamforming may be performed in the frequency domain or time domain. Beamforming may be performed in the frequency domain by (1) multiplying K spatially processed symbols z_i(0) through z_i(K−1) for each transmit antenna i with K weights b_i(0) through b_i(K−1) for that antenna to obtain K transmit symbols and (2) performing OFDM modulation on the K transmit symbols for each transmit antenna i to obtain K time-domain samples for that antenna. Equivalently, beamforming may be performed in the time domain by (1) performing a K-point IDFT on the K spatially processed symbols for each transmit antenna i to obtain K time-domain samples for that transmit antenna and (2) performing a circular convolution of the K time-domain samples for each transmit antenna i with the impulse response g_i(n) for that antenna.

FIG. 4 shows a TX spatial processor 220a that performs beamforming in the frequency domain and is an embodiment of TX spatial processor 220 at transmitting entity 210. TX spatial processor 220a includes a spatial processor 420 and a beamformer 430. Spatial processor 420 performs spatial processing on the data symbols s(k) for each subband k with the eigenmode matrix E(k), the steering matrix V(k), or the identity matrix I and provides spatially processed symbols z(k) for that subband. Beamformer 430 multiplies the spatially processed symbols z(k) for each subband k with the beamforming matrix B(k) to obtain the transmit symbols x(k) for that subband. Modulator 230 performs OFDM modulation on the transmit symbols for each transmit antenna i to obtain a stream of OFDM symbols for that antenna.

FIG. 5 shows an embodiment of spatial processor 420 and beamformer 430 within TX spatial processor 220a. Spatial processor 420 includes K subband spatial processors 520a through 520k for the K subbands and a multiplexer (MUX) 522. Each spatial processor 520 receives the symbols s₀(k) through s_T-1(k) in the vector s(k) for its subband, performs spatial processing on the data symbols with E(k), V(k), or I, and provides spatially processed symbols z₀(k) through z_T-1(k) in the vector z(k) for its subband. Multiplexer 522 receives the spatially processed symbols for all K subbands from spatial processors 520a through 520k and provides these symbols to the proper subbands and transmit antennas.

Beamformer 430 includes T multiplier sets 528a through 528t for the T transmit antennas. For each symbol period, each multiplier set 528 receives the K spatially processed symbols z_i(0) through z_i(K−1) for its transmit antenna i, multiplies these symbols with K weights b_i(0) through b_i(K−1) for transmit antenna i, and provides K transmit symbols x_i(0) through x_i(K−1) for transmit antenna i. For each symbol period, beamformer 430 provides T sets of K transmit symbols for the T transmit antennas.

Modulator 230 includes T OFDM modulator 530a through 530t for the T transmit antennas. Each OFDM modulator 530 receives the transmit symbols x_i(0) through x_i(K−1) for its transmit antenna i, performs OFDM modulation on the transmit symbols, and provides an OFDM symbol for transmit antenna i for each symbol period.

FIG. 6 shows a block diagram of OFDM modulator 530x, which may be used for each of OFDM modulators 530a through 530t in FIG. 5. In each OFDM symbol period, one transmit symbol may be sent on each subband. (A signal value of zero, which is called a zero symbol period, is usually provided for each unused subband.) An IDFT unit 632 receives K transmit symbols for the K subbands in each OFDM symbol period, transforms the K transmit symbols to the time domain with a K-point IDFT, and provides a “transformed” symbol that contains K time-domain samples. Each sample is a complex-value to be transmitted in one sample period. A parallel-to-serial (P/S) converter 634 serializes the K samples for each transformed symbol. A cyclic prefix generator 436 then repeats a portion (or C samples) of each transformed symbol to form an OFDM symbol that contains K+C samples. The cyclic prefix is used to combat inter-symbol interference (ISI) caused by frequency selective fading, which is a frequency response that varies across the overall system bandwidth. An OFDM symbol period (which is also referred to herein as simply a “symbol period”) is the duration of one OFDM symbol and is equal to K+C sample periods.

FIG. 7 shows a TX spatial processor 220b that performs beamforming in the time domain and is another embodiment of TX spatial processor 220 at transmitting entity 210. TX spatial processor 220b includes spatial processor 420 and a beamformer 440. Spatial processor 420 performs spatial processing on the data symbols s(k) for each subband k and provides spatially processed symbols z(k) for that subband. Modulator 230 performs OFDM modulation on the spatially processed symbols for each antenna i and provides a stream of time-domain samples for that antenna. Beamformer 440 performs beamforming in the time-domain by either circularly shifting or linearly delaying the time-domain samples for each transmit antenna i.

FIG. 8A shows a block diagram of modulator 230 and a beamformer 440a, which is one embodiment of beamformer 440 in FIG. 7. Modulator 230 includes T OFDM modulators for the T transmit antennas. Each OFDM modulator includes IDFT unit 632, P/S converter 634, and cyclic prefix generator 636, as shown in FIG. 6. The OFDM modulator for each transmit antenna i receives K spatially processed symbols z_i(0) through z_i(K−1) for the K subbands in each symbol period. Within the OFDM modulator, IDFT unit 632 performs a K-point IDFT on the K spatially processed symbols and provides K time-domain samples. P/S converter 634 serializes the K time-domain samples.

Beamformer 440a includes T circular shift units 842a through 842t for the T transmit antennas. Shift unit 842 for each transmit antenna i receives the K time-domain samples from P/S converter 634 for transmit antenna i, performs a circular shift of the K time-domain samples by i samples, and provides a circular-shifted transformed symbol {z_i′(n)} containing K samples. In particular, shift unit 842a performs a circular shift of zero sample on the transformed symbol {z₀′(n)} for transmit antenna 234a, shift unit 842b performs a circular shift of one sample on the transformed symbol {z₁′(n)} for transmit antenna 234b, and so on, and shift unit 842t performs a circular shift of (T−1) samples on the transformed symbol {z_T-1′(n)} for transmit antenna 234t. T cyclic prefix generators 636a through 636t receive T the circularly-shifted transformed symbols from shift units 842a through 842t, respectively. Each cyclic prefix generator 636 appends a C-sample cyclic prefix to its circularly-shifted transformed symbol {z_i′(n)} and provides an OFDM symbol {x_i(n)} containing (K+C) samples.

FIG. 8B shows a timing diagram for the T transmissions from the T transmit antennas for the embodiment shown in FIG. 8A. T different transformed symbols are generated for the T transmit antennas from T different sets of spatially processed symbols, as shown in FIG. 8A. The T transformed symbols are then circularly shifted by different amounts for the T transmit antennas. A cyclic prefix is appended to each circularly-shifted transformed symbol in the normal manner. The T different OFDM symbols are sent from the T transmit antennas at the same time.

FIG. 9A shows a block diagram of modulator 230 and a beamformer 440b, which is another embodiment of beamformer 440 in FIG. 7. Each OFDM modulator performs OFDM modulation on the spatially processed symbols for its transmit antenna and provides a stream of OFDM symbol {x_i′(n)} for its transmit antenna. Beamformer 440b includes T digital delay units 844a through 844t for the T transmit antennas. Each delay unit 844 receives the OFDM symbol for its transmit antenna i from the associated OFDM modulator and delays the OFDM symbol by an amount determined by transmit antenna i. In particular, delay unit 844a for transmit antenna 234a delays its OFDM symbol {x₀′(n)} by zero sample period, delay unit 844b for transmit antenna 234b delays its OFDM symbol {x′₁(n)} by one sample period, and so on, and delay unit 844t for transmit antenna 234t delays its OFDM symbol {x_T-1′(n)} by T−1 sample periods.

The T different delays may also be provided in the analog domain by transmitter units 232a through 232t. For example, transmitter unit 232a may delay its modulated signal by zero sample period, transmitter unit 232b may delay its modulated signal by one sample period (or T_sam seconds), and so on, and transmitter unit 232t may delay its modulated signal by (T−1) sample periods (or (T−1)·T_sam seconds). A sample period is equal to T_sam=1/[BW·(K+C)], where BW is the overall bandwidth of the system in Hertz.

FIG. 9B shows a timing diagram for the T transmissions from the T transmit antennas for the embodiments shown in FIG. 9A. T different transformed symbols are generated for the T transmit antennas, as shown in FIG. 9A. The OFDM symbol sent from each transmit antenna is delayed by a different amount.

For the embodiments shown in equations (12) through (14) and in FIGS. 8A and 9A, the delays for the T transmit antennas are in integer numbers of sample periods, or i sample periods for transmit antenna i. Other integer phase shifts, instead of i for transmit antenna i, may also be used for antenna i. Phase slopes that result in non-integer delays for the T transmit antennas

$\left(e.g.,g\left(i\right)=e^{-j\pi \frac{i}{L}},\mathrm{for}L>1\right)$ may also be implemented. For example, the time-domain samples from each P/S converter 634 in FIG. 8A may be up-sampled to a higher rate (e.g., with a period of T_upsam=T_sam/L). The higher rate samples may then be circularly shifted by the associated shift unit 842 by integer numbers of the higher rate sample period, T_upsam, where T_upsam<T_sam. Alternatively, each transmitter unit 232 may provide analog delays in integer numbers of T_upsam(instead of T_sam). In general, any amounts of circular or linear delay may be used for the T transmit antennas. The delays for the T transmit antennas should be unique so that no two antennas have the same delay. In the frequency domain, this corresponds to a different phase characteristic for the beamformer across the K subbands.

When the number of transmit antennas is less than the cyclic prefix length (or T<C), the cyclic prefix appended to each OFDM symbol makes a linear delay by each delay units 844 appears like a circular rotation for the circular convolution with the time-domain impulse response g_i(n). The weights as defined in equation (12) may thus be implemented by a time delay of i sample periods for each transmit antenna i, as shown in FIGS. 9A and 9B. However, as shown in FIG. 9B, the T OFDM symbols are transmitted from the T transmit antennas at different delays, which reduces the effectiveness of the cyclic prefix to protect against multipath delay.

Equations (11) and (12) represent a function that provides linearly changing phase shifts across the K subbands for each transmit antenna. The application of linearly changing phase shifts to symbols in the frequency domain may be achieved by either circularly shifting or delaying the corresponding time-domain samples, as described above. In general, the phase shifts across the K subbands for each transmit antenna may be changed in a continuous manner using any function so that the beams are varied in a continuous instead of abrupt manner across the subbands. A linear function of phase shifts is just one example of a continuous function. For a continuous function, an arbitrarily small change in the function input produces an arbitrarily small change in the function output. Some other exemplary continuous functions include a quadratic function, a cubic function, a parabolic function, and so on. The continuous change ensures that the performance of receiving entities that rely on some amounts of correlation across the subbands (e.g., to simplify channel estimation) is not degraded.

The embodiments shown in FIGS. 8A and 9A illustrate some of the ways in which beamforming may be performed in the time domain for continuous beamforming. In general, the beamforming may be performed in various manners and at various locations within the transmitting entity. The beamforming may be performed in the time-domain or the frequency-domain, using digital circuitry or analog circuitry, prior to or after the OFDM modulation, and so on.

The transmitting entity may selectively perform beamforming so that beamforming is either enabled or disabled. The decision to either apply or disable beamforming may be made based on various factors such as, for example, the channel condition. If the transmitting entity performs continuous beamforming, or if the receiving entity performs channel estimation without relying on correlation between subbands, then the receiving entity may not need to be aware of whether or not beamforming is being applied.

The transmitting entity may adaptively perform beamforming so that beamforming is adjusted in some manner over time. In one embodiment, the transmitting entity may enable or disable beamforming based on channel condition, feedback from the receiving entity, and/or some other factors. For example, the transmitting entity may apply beamforming if the channel is flat fading with unit magnitude complex channel gains that may add to zero or a low value for each subband at a receiving entity.

In another embodiment, the transmitting entity may adjust beamforming in a predetermined or pseudo-random manner. For time-domain beamforming, the amounts of delay for the T transmit antennas may be varied for each time interval, which may correspond to one symbol period, multiple symbol periods, the time duration between consecutive transmissions of a MIMO pilot (described below), and so on. For example, the transmitting entity may apply delays of {0, 1, 2, . . . , T−1} sample periods to the T transmit antennas in one time interval, then delays of {0, 0, 0, . . . , 0} sample periods to the T transmit antennas in the next time interval, then delays of {0, 2, 4, . . . , 2(T−1)} sample periods to the T transmit antennas in the following time interval, and so on. The transmitting entity may also cycle through the delays in a base set in different time intervals. For example, the transmitting entity may apply delays of {0, 1, 2, . . . , T−1} sample periods to the T transmit antennas in one time interval, then delays of {T−1, 0, 1, . . . , T−2} sample periods to the T transmit antennas in the next time interval, then delays of {T−2, T−1, 0, . . . , T−3} sample periods to the T transmit antennas in the following time interval, and so on. The transmitting entity may also apply delays in different orders in different time intervals. For example, the transmitting entity may apply delays of {0, 1, 2, . . . , T−1} sample periods to the T transmit antennas in one time interval, then delays of {2, 1, T−1, . . . , 0} sample periods to the T transmit antennas in the next time interval, then delays of {1, T−1, 0, . . . , 2} sample periods to the T transmit antennas in the following time interval, and so on. The transmitting entity may also apply fractional (e.g., 0.5, 1.5) sample periods of delay to any given transmit antenna.

If the receiving entity is unaware that beamforming is being performed, then the transmitting entity may perform beamforming in the same manner across all symbol periods in each data and pilot transmission interval (e.g., each frame). A data and pilot transmission interval is a time interval in which data as well as a pilot used to recover the data are transmitted. For example, the transmitting entity may use the same set of beamforming matrices B(k) for the K subbands or apply the same set of delays to the T transmit antennas for all symbol periods in each data and pilot transmission interval. This allows the receiving entity to estimate an “effective” MIMO channel response (with beamforming) based on a received MIMO pilot and to perform receiver spatial processing on received symbols for the data and pilot transmission interval with the effective MIMO channel response estimate, as described below.

If the receiving entity is aware of the beamforming being performed, then the transmitting entity may adjust the beamforming across the symbol periods in each data and pilot transmission interval. For example, the transmitting entity may use different sets of beamforming matrices B(k) or apply different sets of delays in different symbol periods. The receiving entity may estimate an initial effective MIMO channel response based on a received MIMO pilot, determine the effective MIMO channel response for each subsequent symbol period t based on the initial effective MIMO channel response estimate and knowledge of the beamforming being applied in symbol period t, and perform receiver spatial processing on received symbols for symbol period t with the effective MIMO channel response estimate for symbol period t.

Receiver Spatial Processing

For data transmission with eigensteering and beamforming, the receiving entity obtains R received symbols from the R receive antennas for each subband k, which may be expressed as:r_bes(k)=H(k)·B(k)·E(k)·s(k)+n(k),=H_eff^bes(k)·s(k)+n(k),  Eq. (15)where

r _bes(k) is a vector with R received symbols for subband k;

n(k) is a noise vector for subband k; and

H _eff ^bes(k) is an “effective” channel response matrix observed by data vector s(k) with eigensteering and beamforming, which is:H_eff^bes(k)=H(k)·B(k)·E(k).  Eq. (16)For simplicity, the noise is assumed to be additive white Gaussian noise (AWGN) with a zero mean vector and a covariance matrix of φ_nn=σ²·I, where σ² is the variance of the noise.

The receiving entity can recover the data symbols sent by the transmitting entity using various receiver processing techniques such as a minimum mean square error (MMSE) technique and a channel correlation matrix inversion (CCMI) technique (which is also commonly called a zero-forcing technique).

For the MMSE technique, the receiving entity may derive a spatial filter matrix M_mmse^bes(k) for each subband k, as follows:M_mmse^bes(k)=[H_eff^besH(k)·H_eff^bes(k)+σ²·I]⁻¹·H_eff^besH(k).  Eq. (17)The spatial filter matrix M_mmse^bes(k) minimizes the mean square error between the symbol estimates from the spatial filter and the data symbols in s(k).

The receiving entity may perform MMSE spatial processing for each subband k, as follows:

$\begin{array}{cc}\begin{array}{c}{\widehat{\underset{\_}{s}}}_{\mathrm{mmse}}^{\mathrm{bes}}\left(k\right)={\underset{\_}{D}}_{\mathrm{mmse}}^{\mathrm{bes}}\left(k\right)·{\underset{\_}{M}}_{\mathrm{mmse}}^{\mathrm{bes}}\left(k\right)·{\underset{\_}{r}}_{\mathrm{bes}}\left(k\right),\\ ={\underset{\_}{D}}_{\mathrm{mmse}}^{\mathrm{bes}}\left(k\right)·{\underset{\_}{M}}_{\mathrm{mmse}}^{\mathrm{bes}}\left(k\right)·\left[{\underset{\_}{H}}_{\mathrm{eff}}^{\mathrm{bes}}\left(k\right)·\underset{\_}{s}\left(k\right)+\underset{\_}{n}\left(k\right)\right],\\ \approx \underset{\_}{s}\left(k\right)+{\underset{\_}{n}}_{\mathrm{mmse}}^{\mathrm{bes}}\left(k\right),\end{array}& \mathrm{Eq}.\left(18\right)\end{array}$ where D_mmse^bes(k)=[diag [M_mmse^bes(k)·H_eff^bes(k)]]⁻¹; and

n _mmse ^bes(k) is the MMSE filtered noise.

The symbol estimates from the spatial filter M_mmse^bes(k) are unnormalized estimates of the data symbols. The multiplication with the scaling matrix D_mmse^bes(k) provides normalized estimates of the data symbols.

Eigensteering attempts to send data on the eigenmodes of H(k). However, a data transmission with eigensteering may not be completely orthogonal due to, for example, an imperfect estimate of H(k), error in the eigenvalue decomposition, finite arithmetic precision, and so on. The MMSE technique can account for (or “clean up”) loss of orthogonality in the data transmission with eigensteering.

For the CCMI technique, the receiving entity may derive a spatial filter matrix M_ccmi^bes(k) for each subband k, as follows:M_ccmi^bes(k)=[H_eff^besH(k)·H_eff^bes(k)]⁻¹·H_eff^besH(k).  Eq. (19)

The receiving entity may perform CCMI spatial processing for each subband k, as follows:

$\begin{array}{cc}\begin{array}{c}{\widehat{\underset{\_}{s}}}_{\mathrm{ccmi}}^{\mathrm{bes}}\left(k\right)={\underset{\_}{M}}_{\mathrm{ccmi}}^{\mathrm{bes}}\left(k\right)·{\underset{\_}{r}}_{b\mathrm{es}}\left(k\right),\\ ={\left[{\underset{\_}{H}}_{\mathrm{eff}}^{{\mathrm{bes}}^H}\left(k\right)·{\underset{\_}{H}}_{\mathrm{eff}}^{\mathrm{bes}}\left(k\right)\right]}^{-1}·{\underset{\_}{H}}_{\mathrm{eff}}^{{\mathrm{bes}}^H}\left(k\right)·\\ \left[{\underset{\_}{H}}_{\mathrm{eff}}^{\mathrm{bes}}\left(k\right)·\underset{\_}{s}\left(k\right)+\underset{\_}{n}\left(k\right)\right],\\ =\underset{\_}{s}\left(k\right)+{\underset{\_}{n}}_{\mathrm{ccmi}}^{\mathrm{bes}}\left(k\right),\end{array}& \mathrm{Eq}.\left(20\right)\end{array}$ where n_ccmi^bes(k) is the CCMI filtered noise. The CCMI technique may amplify the noise due to the structure of R_eff^bes(k)=H_eff^besH(k)·H_eff^bes(k).

The receiving entity may perform spatial processing for the other operating modes in similar manner, albeit with different effective channel response matrices and different spatial filter matrices. Table 1 summarizes the spatial processing at the transmitting entity for the various operating modes and the effective MIMO channel for each operating mode. For clarity, the index “(k)” for subband is not shown in Table 1. Beamforming may be performed in the frequency domain, as shown Table 1. Linear continuous beamforming may also be performed in the time domain, as described above. In this case, the beamforming matrix B is omitted from the transmit symbol vector x but is still present in the effective MIMO channel response.

TABLE 1
	Transmitter	Effective Channel
No	Eigensteering	x es = E · s	H eff es = H · E
Beam-	Matrix Steering	x ss = V · s	H eff ss = H · V
forming	No Spatial Processing	x ns = s	H eff ns = H
Beam-	Eigensteering	x bes = B · E · s	H eff bes = H · B · E
forming	Matrix Steering	x bss = B · V · s	H eff bss = H · B · V
	No Spatial Processing	x bns = B · s	H eff bns = H · B

In general, the receiving entity may derived an MMSE spatial filter matrix M_mmse^x(k) for each subband k, as follows:M_mmse^x(k)=[H_eff^xH(k)·H_eff^x(k)+σ²·I]⁻¹·H_eff^xH(k),  Eq. (21)where the superscript “x” denotes the operating mode and may be equal to “es” for eigensteering without beamforming, “ss” for matrix steering without beamforming, “ns” for no spatial processing and no beamforming, “bes” for eigensteering with beamforming, “bss” for matrix steering with beamforming, or “bns” for beamforming only. The MMSE spatial filter matrix M_mmse^x(k) may be derived in the same manner for all operating modes, albeit with different effective channel response matrices H_eff^es(k), H_eff^ss(k), H_eff^ns(k), H_eff^bes(k), H_eff^bss(k), and H_eff^bns(k). The MMSE receiver spatial processing may also be performed in the same manner for all operating modes, albeit with the MMSE spatial filter matrices being derived with different effective channel response matrices. An MMSE-based receiver may thus support all operating modes using the same MMSE spatial processing. In equation (21), the term σ²·I may be replaced with the covariance matrix φ_nn of the noise, if known.

The receiving entity may also derived a CCMI spatial filter matrix M_ccmi^x(k) for each subband k, as follows:M_ccmi^x(k)=[H_eff^xH(k)·H_eff^x(k)]⁻¹·H_eff^xH(k).  Eq (22)Again, the receiving entity may derive the CCMI spatial filter matrix in the same manner for all operating modes, albeit with different effective channel response matrices. The receiving entity may also apply the CCMI spatial filter matrices in the same manner for all operating modes.

The receiving entity may utilize other receiver spatial processing techniques to recover the data symbols, and this is within the scope of the invention.

Pilot Transmission

The transmitting entity may transmit a pilot to allow the receiving entity to estimate the actual or effective MIMO channel response. The pilot may be transmitted in various manners. For example, the transmitting entity may transmit an unsteered MIMO pilot, a steered MIMO pilot, a spread MIMO pilot, and so on. A MIMO pilot is a pilot comprised of multiple pilot transmissions sent from the T transmit antennas. An unsteered MIMO pilot is comprised of up to T pilot transmissions sent from the T transmit antennas, one pilot transmission from each antenna. A steered MIMO pilot is comprised of up to S pilot transmissions sent on the S orthogonal spatial channels. A spread MIMO pilot is comprised of up to S pilot transmissions sent on the S spatial channels with matrix steering.

For a MIMO pilot, each of the multiple pilot transmissions is identifiable by the receiving entity. This may be achieved by:

• • 1. Apply a different orthogonal sequence to each pilot transmission using code division multiplexing (CDM),
  • 2. Send the multiple pilot transmissions in different symbol periods using time division multiplexing (TDM), and/or
  • 3. Send the multiple pilot transmissions on different subbands using frequency division multiplexing (FDM).For FDM, a different set of subbands may be used for each of the multiple pilot transmissions. The subbands used for each pilot transmission may be cycled such that the pilot transmission eventually observes all K subbands. A MIMO pilot may be sent with full transmit power for each transmit antenna using CDM or FDM, which is desirable. A MIMO pilot may also be sent using any combination of CDM, FDM, and TDM.

For an unsteered MIMO pilot, the transmitting entity may perform spatial processing for each subband k used for pilot transmission as follows:z_ns,np(k,t)=W(t)·p(k),  Eq. (23)where

p(k) is a vector of pilot symbols to be sent on subband k;

W(t) is a diagonal Walsh matrix for symbol period t; and

z _ns,mp(k) is a vector of spatially processed symbols for the unsteered MIMO pilot for subband k in symbol period t.

Different pilot symbols may be sent from the T transmit antennas, as shown in equation (23). Alternatively, the same pilot symbol may also be used for all transmit antennas, in which case the Walsh matrix is simply a Walsh vector.

If T=4, then the four transmit antennas may be assigned 4-symbol Walsh sequences W₁=1, 1, 1, 1, W₂=1, −1, 1, −1, W₃=1, 1, −1, −1, and W₄=1, −1, −1, 1 for the MIMO pilot. The four symbols of Walsh sequence W_j are applied to the pilot transmission from transmit antenna j in four symbol periods. W(1) contains the first element of the four Walsh sequences along its diagonal, W(2) contains the second element of the four Walsh sequences, W(3) contains the third element of the four Walsh sequences, and W(4) contains the fourth element of the four Walsh sequences. The j-th Walsh sequence W_j for transmit antenna j is thus carried as the j-th diagonal element of all the Walsh matrices. The four Walsh matrices may be used in four symbol periods to transmit the unsteered MIMO pilot.

The transmitting entity further processes the vector z_ns,mp(k,t) for either beamforming or no beamforming, e.g., in the same manner as the data vector s(k), to obtain a transmit vector for the unsteered MIMO pilot. The transmitting entity may transmit the unsteered MIMO pilot over T symbol periods by using one Walsh matrix W(t) for each symbol period.

For an unsteered MIMO pilot without beamforming, the receiving entity obtains received pilot symbols for each subband k used for pilot transmission, as follows:r_ns,mp(k,t)=H(k)·W(t)·p(k)+n(k).  Eq. (24)The MIMO channel and noise is assumed to be static over the time during which the unsteered MIMO pilot is transmitted. The receiving entity obtains T vectors r_ns,mp(k,1) through r_ns,mp(k,T) for T-symbol Walsh sequences used for the unsteered MIMO pilot.

The receiving entity may estimate the actual MIMO channel response H(k) based on the received unsteered MIMO pilot without beamforming. Each column j of H(k) is associated with a respective Walsh sequence W_j. The receiving entity may obtain h_i,j(k), which is the i-th element of the j-th column of H(k) by (1) multiplying the i-th element of r_ns,mp(k,1) through r_ns,mp(k,T) by the T chips of the Walsh sequence W_j, (2) removing the modulation used for pilot symbol p_j(k), which is the j-th element of p(k), and (3) accumulating the T resultant elements to obtain h_i,j(k). The process may be repeated for each element of H(k). The receiving entity may then use H(k) to derive the effective MIMO channel response H_eff^ss(k) or H_eff^es(k), which may be used for receiver spatial processing.

For an unsteered MIMO pilot with beamforming, the receiving entity obtains received pilot symbols for each subband k used for pilot transmission, as follows:r_bns,mp(k,t)=H(k)·B(k)·W(t)·p(k)+n(k).  Eq. (25)The receiving entity may perform similar processing on the received unsteered MIMO pilot with beamforming to obtain H_eff^bns(k) or H_eff^bss(k).

For a steered MIMO pilot, the transmitting entity may perform spatial processing for each subband k used for pilot transmission as follows:z_es,mp(k,t)=E(k)·W(t)·p(k),  Eq. (26)where z_es,mp(k,t) is a vector of spatially processed symbols for the steered MIMO pilot for subband k in symbol period t. For simplicity, E(k) is assumed to be static over the time during which the steered MIMO pilot is transmitted, and is thus not a function of symbol period t. The transmitter may further process the vector z_es,mp(k,t) for either beamforming or no beamforming and may then transmit the steered MIMO pilot.

For a steered MIMO pilot without beamforming, the receiving entity obtains received pilot symbols for each subband k used for pilot transmission, as follows:r_es,mp(k,t)=H(k)·E(k)·W(t)·p(k)+n(k).  Eq. (27)For a steered MIMO pilot with beamforming, the receiving entity obtains received pilot symbols for each subband k used for pilot transmission, as follows:r_bes,mp(k,t)=H(k)·B(k)·E(k)·W(t)·p(k)+n(k).  Eq. (28)The receiving entity may estimate H_eff^es(k) based on r_es,mp(k,n) and may estimate the H_eff^bes(k) based on r_bes,mp(k,n), in similar manner as described above for H(k).

For a spread MIMO pilot, the transmitting entity may perform spatial processing for each subband k used for pilot transmission as follows:z_ss,mp(k,t)=V(k)·W(t)·p(k),  Eq. (29)where z_ss,mp(k,t) is a vector of spatially processed symbols for the spread MIMO pilot for subband k. The transmitter may further process the vector z_ss,mp(k,t) for either beamforming or no beamforming, and may then transmit the resultant MIMO pilot.

The receiving entity may estimate H_eff^ss(k) based on a received spread MIMO pilot without beamforming and may estimate H_eff^bss(k) based on a received spread MIMO pilot with beamforming. The receiving entity may then derive the effective MIMO channel response H_eff^ns(k) or H_eff^bns(k), which may be used for receiver spatial processing.

Steering Matrix

A set of steering matrices may be generated and used for matrix steering. These steering matrices may be denoted as {V}, or V(i) for i=1 . . . L, where L may be any integer greater than one. Each steering matrix V(i) should be a unitary matrix. This condition ensures that the T data symbols transmitted simultaneously using V(i) have the same power and are orthogonal to one another after the matrix steering with V(i).

The set of L steering matrices may be generated in various manners. For example, the L steering matrices may be generated based on a unitary base matrix and a set of scalars. The base matrix may be used as one of the L steering matrices. The other L−1 steering matrices may be generated by multiplying the rows of the base matrix with different combinations of scalars. Each scalar may be any real or complex value. The scalars are selected to have unit magnitude so that steering matrices generated with these scalars are unitary matrices.

The base matrix may be a Walsh matrix. A 2×2 Walsh matrix W_2×2 and a larger size Walsh matrix W_2N×2N may be expressed as:

$\begin{array}{cc}{\underset{\_}{W}}_{2\times 2}=\left[\begin{array}{cc}1& 1\\ 1& -1\end{array}\right]\mathrm{and}{\underset{\_}{W}}_{2N\times 2N}=\left[\begin{array}{cc}{\underset{\_}{W}}_{N\times N}& {\underset{\_}{W}}_{N\times N}\\ {\underset{\_}{W}}_{N\times N}& -{\underset{\_}{W}}_{N\times N}\end{array}\right].& \mathrm{Eq}.\left(30\right)\end{array}$ Walsh matrices have dimensions that are powers of two (e.g., 2, 4, 8, and so on).

The base matrix may also be a Fourier matrix. For an N×N Fourier matrix D_N×N, the elements d_n,m of D_N×N may be expressed as:

$\begin{array}{cc}{d}_{n,m}=e^{-j2{\pi }^{\frac{n·m}{N}}},\mathrm{for}n=0,\dots ,N-1\mathrm{and}m=0,\dots ,N-1.& \mathrm{Eq}.\left(31\right)\end{array}$ Fourier matrices of any square dimension (e.g., 2, 3, 4, 5, and so on) may be formed. Other matrices may also be used as the base matrix.

For an N×N base matrix, each of rows 2 through N of the base matrix may be independently multiplied with one of Q different possible scalars. Q^N-1 different steering matrices may be obtained from Q^N-1 different permutations of the Q scalars for N−1 rows. For example, each of rows 2 through N may be independently multiplied with a scalar of +1, −1, +j, or −j, where j=√{square root over (−1)}. In general, each row of the base matrix may be multiplied with any scalar having the form e^jθ, where θ may be any phase value. Each element of a scalar-multiplied N×N base matrix is further scaled by 1/√{square root over (N)} to obtain an N×N steering matrix having unit power for each column.

Steering matrices derived based on a Walsh matrix (or a 4×4 Fourier matrix) have certain desirable properties. If the rows of the Walsh matrix are multiplied with scalars of ±1 and ±j, then each element of a resultant steering matrix V(i) belongs in a set composed of {+1, −1, +j, −j}. In this case, the multiplication of an element of another matrix with an element of V(i) may be performed with just bit manipulation.

The data transmission techniques described herein may be used for various wireless systems. These techniques may also be used for the downlink (or forward link) as well as the uplink (or reverse link).

Continuous beamforming with or without matrix steering may be used in various manners. For example, a transmitting entity (e.g., an access point or a user terminal) may use continuous beamforming to transmit to a receiving entity (e.g., another access point or user terminal) when accurate information about the wireless channel is not available. Accurate channel information may not be available due to various reasons such as, for example, a feedback channel that is corrupted, a system that is poorly calibrated, the channel conditions changing too rapidly for the transmitting entity to use/adjust beam steering on time (e.g., due to the transmitting and/or receiving entity moving at a high velocity), and so on.

Continuous beamforming may also be used for various applications in a wireless system. In one application, broadcast channels in the system may be transmitted using continuous beamforming, as described above. The use of continuous beamforming allows wireless devices in the system to receive the broadcast channels with improved reliability, thereby increasing the range of the broadcast channels. In another application, a paging channel is transmitted using continuous beamforming. Again, improved reliability and/or greater coverage may be achieved for the paging channel via the use of continuous beamforming. In yet another application, an 802.11a access point uses continuous beamforming to improve the performance of user terminals under its coverage area.

The transmission techniques described herein may be implemented by various means. For example, these techniques may be implemented in hardware, software, or a combination thereof. For a hardware implementation, the processing units at a transmitting entity may be implemented within one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, micro-controllers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof. The processing units at a receiving entity may also be implemented with one or more ASICs, DSPs, and so on.

For a software implementation, some of the processing may be implemented with modules (e.g., procedures, functions, and so on) that perform the functions described herein. The software codes may be stored in a memory unit (e.g., memory unit 242 or 282 in FIG. 2) and executed by a processor (e.g., controller 240 or 280). The memory unit may be implemented within the processor or external to the processor, in which case it can be communicatively coupled to the processor via various means as is known in the art.

Headings are included herein for reference and to aid in locating certain sections. These headings are not intended to limit the scope of the concepts described therein under, and these concepts may have applicability in other sections throughout the entire specification.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for receiving data in a multiple-input and multiple-output communication system, comprising: receiving a multiple-input and multiple-output pilot;determining an effective multiple-input and multiple-output channel response based on estimated initial effective multiple-input and multiple-output channel response that is based on the multiple-input and multiple-output pilot; andperforming receiver spatial processing on received symbols for a symbol period with the determined effective multiple-input and multiple-output channel response to recover data symbols.

2. The method of claim 1, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with an eigenmode matrix.

3. The method of claim 1, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with a steering matrix.

4. The method of claim 1, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with an identity matrix.

5. The method of claim 1, wherein the data received was transmitted using beamforming in the frequency domain.

6. The method of claim 1, wherein the data received was transmitted using beamforming in the time domain.

7. The method of claim 1, wherein the data received was transmitted without beamforming.

8. The method of claim 1, wherein performing receiver spatial processing on received symbols comprises performing minimum mean square error spatial processing for each subband.

9. The method of claim 8, wherein performing minimum mean square error spatial processing comprises: deriving a spatial filter matrix for each subband to obtain symbol estimates; andmultiplying the symbol estimates with a scaling matrix to obtain normalized estimates of the data symbols.

10. The method of claim 1, wherein performing receiver spatial processing on received symbols comprises performing channel correlation matrix inversion for each subband.

11. A wireless device in a multiple-input and multiple-output communication system, comprising: a processor;memory in electronic communication with the processor;instructions stored in the memory, the instructions being executable by the processor to: receive a multiple-input and multiple-output pilot;determine an effective multiple-input and multiple-output channel response based on an estimated initial effective multiple-input and multiple-output channel response that is based on the multiple-input and multiple-output pilot; andperform receiver spatial processing on received symbols for a symbol period with the determined effective multiple-input and multiple-output channel response to recover data symbols.

12. The wireless device of claim 11, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with an eigenmode matrix.

13. The wireless device of claim 11, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with a steering matrix.

14. The wireless device of claim 11, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with an identity matrix.

15. The wireless device of claim 11, wherein the data received was transmitted using beamforming in the frequency domain.

16. The wireless device of claim 11, wherein the data received was transmitted using beamforming in the time domain.

17. The wireless device of claim 11, wherein the data received was transmitted without beamforming.

18. The wireless device of claim 11, wherein the instructions executable to perform receiver spatial processing on received symbols comprises instructions executable to perform minimum mean square error spatial processing for each subband.

19. The wireless device of claim 18, wherein the instructions executable to perform minimum mean square error spatial processing comprises instructions executable to: derive a spatial filter matrix for each subband to obtain symbol estimates; andmultiply the symbol estimates with a scaling matrix to obtain normalized estimates of the data symbols.

20. The wireless device of claim 11, wherein the instructions executable to perform receiver spatial processing on received symbols comprises instructions executable to perform channel correlation matrix inversion for each subband.

21. An apparatus configured for receiving data in a multiple-input and multiple-output communication system, comprising: means for receiving a multiple-input and multiple-output pilot;means for determining an effective multiple-input and multiple-output channel response based on an estimated initial effective multiple-input and multiple-output channel response that is based on the multiple-input and multiple-output pilot; andmeans for performing receiver spatial processing on received symbols for a symbol period with the determined effective multiple-input and multiple-output channel response to recover data symbols.

22. The apparatus of claim 21, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with an eigenmode matrix.

23. The apparatus of claim 21, wherein the data received was transmitted using spatial processing of the data symbols for each frequency subband with a steering matrix.

24. The apparatus of claim 21, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with an identity matrix.

25. The apparatus of claim 21, wherein the data received was transmitted using beamforming in the frequency domain.

26. The apparatus of claim 21, wherein the data received was transmitted using beamforming in the time domain.

27. The apparatus of claim 21, wherein the data received was transmitted without beamforming.

28. The apparatus of claim 21, wherein performing receiver spatial processing on received symbols comprises performing minimum mean square error spatial processing for each subband.

29. The apparatus of claim 28, wherein the means for performing minimum mean square error spatial processing comprises: means for deriving a spatial filter matrix for each subband to obtain symbol estimates; andmeans for multiplying the symbol estimates with a scaling matrix to obtain normalized estimates of the data symbols.

30. The apparatus of claim 21, wherein performing receiver spatial processing on received symbols comprises performing channel correlation matrix inversion for each subband.

31. A computer-program product for receiving data in a multiple-input and multiple-output communication system, the computer-program product comprising a non-transitory computer-readable medium having instructions thereon, the instructions comprising: code for causing a wireless device to receive a multiple-input and multiple-output pilot;code for causing the wireless device to determine an effective multiple-input and multiple-output channel response based on an estimated initial effective multiple-input and multiple-output channel response that is based on the multiple-input and multiple-output pilot; andcode for causing the wireless device to perform receiver spatial processing on received symbols for a symbol period with the determined effective multiple-input and multiple-output channel response to recover data symbols.

32. The computer-program product of claim 31, wherein the data comprises data symbols for each frequency subband that have been spatially processed with an eigenmode matrix.

33. The computer-program product of claim 31, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with a steering matrix.

34. The computer-program product of claim 31, wherein the data received comprises data symbols for each frequency subband that have been spatially processed with an identity matrix.

35. The computer-program product of claim 31, wherein the data received was transmitted using beamforming in the frequency domain.

36. The computer-program product of claim 31, wherein the data received was transmitted using beamforming in the time domain.

37. The computer-program product of claim 31, wherein the data received was transmitted without beamforming.

38. The computer-program product of claim 31, wherein the code for causing the wireless device to perform receiver spatial processing on received symbols comprises code for causing the wireless device to perform minimum mean square error spatial processing for each subband.

39. The computer-program product of claim 38, wherein the code for causing the wireless device to perform minimum mean square error spatial processing comprises: code for causing the wireless device to derive a spatial filter matrix for each subband to obtain symbol estimates; andcode for causing the wireless device to multiply the symbol estimates with a scaling matrix to obtain normalized estimates of the data symbols.

40. The computer-program product of claim 31, wherein the code for causing the wireless device to perform receiver spatial processing on received symbols comprises performing channel correlation matrix inversion for each subband.