Frequency-independent spatial processing for wideband MISO and MIMO systems

ES 2 525 141 T3

DESCRIPTION

Frequency independent spatial processing for wideband MISO and MIMO systems

Background

I. Field

The present invention relates generally to data communication and more specifically to techniques for performing spatial processing for multiple input single output (MISO) and multiple input multiple output (MIMO) communication systems of broadband.

II. Background

A MIMO system employs multiple (Nt) transmit antennas and multiple (Nr) receive antennas for data transmission, and is denoted as one (Nt, Nr) system. A MIMO channel formed by the Nt transmitting antennas and the Nr receiving antennas can be decomposed into Ns independent channels, where Ns <min {Nt, Nr}. Ns spatial channels can be formed by the Ns independent channels of the MIMO channel, and be used for data transmission.

For a time-dispersive MIMO channel, a signal sent from a given transmitting antenna can reach a given receiving antenna via multiple signal paths (ie, propagation paths). These signal paths can include a straight line path and / or reflected paths, which are created when the transmitted signal is reflected from reflection sources (e.g. buildings, obstructions, etc.) and reaches the receiving antenna via signal paths other than the straight-line path. The signal received at the receiving antenna may therefore include multiple instances (ie, multi-path components) of the signal sent from the transmitting antenna. The MIMO channel delay spread L is the time difference between the earliest and the latest multi-path components arrived (of some certain minimum energy) for all transmit and receive antenna pairs on the MIMO channel.

Time spread in the MIMO channel produces selective frequency fading, which is characterized by a frequency response that varies across the system bandwidth (ie, different channel gains for different frequencies). Multi-path components are associated with different complex channel gains and can add constructively, or destructively, at the receiver. Time dispersion and frequency selective fading are more problematic for a wideband MIMO system with a wide system bandwidth.

Various techniques can be used to combat frequency selectivity in a wideband MIMO channel. For example, a multi-carrier modulation technique, such as orthogonal frequency division multiplexing (OFDM), can be used to divide the system bandwidth into multiple (Nr) orthogonal frequency sub-bands. The wideband MIMO channel can then be viewed as composed of Nr flat-fade narrowband MIMO channels, each of which can be decomposed into Ns spatial channels. The data can then be transmitted on the Ns spatial channels of each of the Nr subbands.

For a MIMO system using OFDM (i.e. a MIMO-OFDM system), the wideband MIMO channel can be characterized with (1) a complex channel gain for each of the Nr subbands of each one of the Nt · Nr transmit / receive antenna pairs (ie, Nr · Nt · Nr channel gains in total) and (2) the lower limit of the noise at the receiver. The channel gains and receiver noise lower limit can then be used to select the data rate (s) for data transmission on the Ns spatial channels of each of the Nr sub-bands. The channel gains can also be used for spatial processing at the receiver and possibly at the transmitter in order to transmit data on the Ns spatial channels of each of the Nr subbands. Thus, for the MIMO-OFDM system, the frequency selectivity can be combated by treating the wideband MIMO channel as Nr flat-fading narrowband MIMO channels, and performing the spatial processing separately for each of narrowband MIMO channels. However, this frequency-dependent spatial processing can greatly increase computational complexity at the transmitter and receiver. In addition, the receiver may need to provide a large amount of feedback information (eg, channel gains) to the transmitter to support frequency-dependent spatial processing.

Document WO 02/078211 discloses obtaining a correlated filter for signal reception.

Therefore, there is a need in technology for techniques to more efficiently perform spatial processing in a wideband MIMO system.

ES 2 525 141 T3

Summary

The preceding problem is solved by the independent claims.

Techniques for performing frequency independent self-guiding in MISO and MIMO systems are provided herein. Self-guiding refers to the spatial processing performed on a data symbol stream with a guide vector in a transmitter, in order to transmit the data symbol stream over a spatial channel of a MISO channel or a MIMO channel. . The MISO channel can be characterized by (1) a sequence of time domain channel impulse response vectors, for a plurality of time delays, or (2) a sequence of channel frequency response vectors from the the frequency, for the Nf subbands. Similarly, the MIMO channel can be characterized by a sequence of channel impulse response matrices, or a sequence of channel frequency response matrices. Self-guiding is independent of frequency in that a guide vector is used for the flow of data symbols, even if the MISO or MIMO channel is time scatter, and regardless of whether self-guiding is performed in the time domain or in the frequency domain. Self-guiding can be performed on one or multiple data symbol streams, with one or multiple guide vectors, to transmit the data symbol stream (s) over one or multiple spatial channels. Various frequency-independent self-guidance schemes are described herein, including main mode self-guidance, multi-mode self-guidance, main path self-guidance, and receiver self-guidance. .

For main mode autoguide and multi-mode autoguide, a correlation matrix is ​​calculated for the MIMO channel, based on the channel response matrices (impulse or frequency) for the channel. MIMO, as described below. The correlation matrix is ​​then decomposed (eg, using self-value decomposition) to obtain Ns frequency independent guide vectors for Ns spatial channels of the MIMO channel. For main mode self-guiding, a stream of data symbols is transmitted on the main spatial channel, or the best one, using the guide vector v_p.m for the best space channel. For multi-modality self-guidance, Nd data symbol streams are transmitted on the Nd best spatial channels, using Nd guide vectors Vmm for these spatial channels, where Ns> Nd> 1 in this case.

For main path self-guiding, a stream of data symbols is transmitted on the main spatial channel for the main propagation path of the MIMO channel, using a frequency independent guide vector Vmp. For this scheme, the energy of each channel impulse response matrix is ​​first determined. The main path is the time delay of the channel impulse response matrix with the highest energy. A correlation matrix of the channel impulse response matrix with the highest energy is calculated and decomposed to obtain the guide vector Vmp for the best spatial channel of the main path. The data symbol stream is transmitted on this spatial channel using the guide vector Vmp.

For receiver self-guidance, a stream of data symbols is guided towards an individual receiving antenna, based on a guide vector and_rx, independent of frequency, obtained for that receiving antenna. The MIMO channel can be seen as composed of Nr MISO channels for the Nr receiving antennas. A correlation matrix can be calculated for each MISO channel, based on its sequence of channel response vectors (impulse or frequency), and decomposed to obtain a guide vector for the main spatial channel of that MISO channel. . Nr frequency independent V.rx guide vectors can be obtained for the Nr MISO channels. Nd data symbol streams can be transmitted using the Nr guide vectors V.rx, where min {Nr, Nt}> Nd> 1 in this case. Each stream of data symbols can be guided toward one, multiple, or all of the receiving antennas. For a MISO system with a receiving antenna, a guide vector is obtained for the single receiving antenna and is used to transmit a stream of data symbols.

For all self-guiding schemes, a correlated filter is obtained for each receiving antenna, based on the guide vector (s) used by the transmitter and the sequence of response vectors (impulse or signal). frequency) channel for the receiving antenna. The stream of symbols received for each receiving antenna is filtered with the correlated filter for that receiving antenna, to obtain one or more sub-streams of filtered symbols. The filtered symbol sub-streams from all the Nr correlated filters for the Nr receive antennas are then combined to obtain Nd detected symbol streams for the Nd data streams sent by the transmitter, where Nd> 1 in this case. Equalization and other post-processing can be performed on the Nd detected symbol streams to obtain Nd recovered symbol streams, which are estimates of the Nd data symbol streams sent by the transmitter.

Various aspects and embodiments of the invention are described in greater detail below.

Brief description of the drawings

The characteristics and nature of the present invention will become more apparent from the detailed description set forth below, when considered in conjunction with the drawings, in which the

ES 2 525 141 T3 equal reference characters identify correspondingly in all their extension, and in which:

FIG. 1 shows a transmitter and a receiver in a MISO system;

FIG. 2 shows a transmission data processor (TX) in the MISO system;

FIGs. 3A, 3B and 3C show three embodiments of a TX spatial processor in the MISO system;

FIGs. 4A, 4B and 4C show three embodiments of a receive spatial processor (RX) in the MISO system;

FIG. 5 shows a block diagram of the receiver in the MISO system;

FIG. 6 shows a transmitter and a receiver in a MIMO system;

FIG. 7 shows a TX data processor in the MIMO system;

FIGs. 8A, 8B and 8C show three embodiments of a TX spatial processor in the MIMO system;

FIGs. 9A to 9F show six embodiments of an RX spatial processor in the MIMO system;

FIG. 10 shows a block diagram of the receiver in the MIMO system;

FIG. 11 shows a process for realizing main mode self-guidance, multi-mode self-guidance and main path self-guidance in the MIMO system; and FIG. 12 shows a process for performing receiver self-guidance in the MISO or MIMO system.

Detailed description

The word "exemplary" is used herein to express "serving as an example, case, or illustration." Any embodiment or design described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments or designs.

The self-guiding techniques described herein can be used for various wireless communication systems, including MISO and single-carrier and multi-carrier MIMO systems. The multiple carriers can be provided by OFDM or some other multi-carrier modulation technique or structure. In the description below, the term "MIMO system" refers generically to both single-carrier and multi-carrier MIMO systems.

For clarity, the following notations are used for the description below. Time domain variables are functions of n and are indicated with italicized text (eg, h (n)). Frequency domain variables are functions of k and are indicated in plain text (eg, h (k)). Vectors are indicated with lowercase, bold, and underlined text (eg, h (n) and h (k)). The matrices are indicated with uppercase, bold and underlined text (p. g., H (n) and H (k)). Three-dimensional matrices are indicated with uppercase, bold, and double underlined text (eg, H and H).

1. MISO system

A time-dispersing MISO channel with Nt transmitting antennas and a single receiving antenna can be characterized by a time-domain channel impulse response matrix H, with dimensions (L + 1) x Nt, where L indicates the The magnitude of the MISO channel delay in symbol periods. The magnitude of a channel's delay is the difference between the earliest and latest resolvable propagation paths in the channel. The - matrix is ​​composed of Nt channel impulse response vectors hj, for J = 1, 2, ..., Nt or, equivalently, L + 1 row vectors h (n), for n = 0, 1,. .. L, which can be expressed as:

<td>w</td><td></td><td>X (0) 4 (0) Λ ^ (0)</td>

<td>Λ (1)</td><td></td><td>40) 40) A and_r(l)</td>

<td>M</td><td></td><td>Μ Μ OM</td>

<td></td><td></td><td>W) 4 (¿) A and_r(^)</td>

, Eq. (1) where the input hj (n), for J = 1, 2, ... Nty n = 0, 1, ... L, is the coupling (that is, the complex gain) between the antenna J of transmitting and receiving antenna for time delay n. Each vector hj, for J = 1, 2, ..., Nt, includes L + 1 complex values ​​for the channel impulse response between the transmitting antenna J and the signaling antenna.

ES 2 525 141 T3 reception. Each row vector h (n), for n = 0, 1, L, includes complex Ntvalues ​​for the channel gains between the transmitting Ntantenas and the receiving antenna for the time delay n.

The channel impulse response can be estimated by the receiver based on pilot symbols sent by the transmitter. The transmitter can "cover" the pilot signal for each transmitting antenna with a single orthogonal sequence assigned to that antenna. Coverage is a process by which a given modulation symbol p (or a set of W symbols with the same value) to be transmitted is multiplied by all the W segments of an orthogonal sequence of W segments, to obtain W covered symbols, which they are then transmitted. Nt orthogonal pilot signals can be obtained with Nt orthogonal sequences for the Nt transmitting antennas. The coverage achieves orthogonality between the Nt pilot signals sent from the Nt transmitting antennas and allows the receiver to distinguish the individual transmitting antennas.

The receiver can "discover" the received pilot symbols with each of the same Nt orthogonal sequences, to estimate the channel impulse response between each of the Nt transmit antennas and the receive antenna. Discovery is a complementary process, whereby W received symbols for W covered symbols are multiplied by the W segments of the same orthogonal sequence of W segments, to obtain W discovered symbols, which are then accumulated to obtain an estimate of the transmitted symbol p . The discovery is performed on L + 1 time delays, to obtain the L + 1 row vectors h (n), for n = 0, 1, L, for the channel impulse response of the MISO channel.

A time dispersing MISO channel can also be characterized by a two-dimensional frequency-domain channel frequency response matrix H, with dimensions Nf x Nt, where Nf is the number of frequency sub-bands and Nf&gt; (L +1). The matrix H is composed of Ntvectors h / of channel frequency response, for j = 1,2, Nt, that is, H = [hi h2 ... h / vr]. Each vector h includes A / ^ values ​​of the frequency domain that can be obtained by performing a discrete Fourier transformation (DFT) of Nf points on the L + 1 values ​​of the time domain of the corresponding vector h, of the matrix H. For each pair of transmitting and receiving antennas, there is thus a one-to-one correspondence between the time domain channel impulse response vector hj and the frequency domain channel frequency response vector h. The matrix H is equivalently composed of Nf row vectors h (k), for k = 1, 2, Nf, that is, H = [h^T(1 hour^r(2) ... ή^Γ(Λ / ρ)]^Γ, where M^r is the transpose of M. Each row vector h (k) includes Nt complex gain values ​​for the frequency response between the Nt transmitting antennas and the receiving antenna for sub-band k. The DFT to obtain h (k) can be expressed as:

a&gt; 0, for k = 1,2 ..... Nf.

Eq. (2)

The transmitter can self-guide over a data symbol stream s (n) prior to transmission, via the MISO channel, in order to achieve a higher received signal to noise ratio (SNR) for the symbol stream of data. Self-guidance can be done using a vector v_m¡_SW frequency independent guidance, which can be obtained to maximize the received SNR, or based on some other criteria.

In one embodiment, the vector v_m¡_SW The guideline for the transmitter is obtained by first computing a matrix R_m/_Sor correlation of dimensions NtxNt, according to the following:

. _{Ec (3)} where h^H is the conjugate transpose of h. The matrix R_m/_Sor it can be seen as an average of L + 1 individual correlation matrices of ft (n) for L + 1 time lags, or Nf individual correlation matrices of h (k) for Nf sub-bands. The individual correlation matrices are given equal weight in equation (3). In another embodiment, the individual correlation matrices may be given different weights in the calculation of R_myso · For example, each individual correlation matrix can be weighted by the energy associated with that matrix, which can be calculated as described below.

The decomposition into eigenvalues ​​of the matrix R_mys₀ Correlation is then performed as follows:

<img file="ES2525141T3_D0001.tif" />

Eq. (4) where

ES 2 525 141 T3

Vmiso is a unit matrix of dimensions Λ /? χ Λ /? whose columns are the eigenvectors of R_mys₀; Y

TO_mSo is a diagonal matrix of dimensions NtxNt whose diagonal entries are the eigenvalues ​​of R_mySW·

A unitary matrix M is characterized by the property M^HM = |, where I is the identity matrix, with ones along the diagonal and zeros in the rest. The Nt auto-vectors of the unit matrix V_m¡_Sor, indicated as v, for / = 1,2, ..., Nt, are therefore orthogonal to each other. Also, the length of each self-vectors equal to one, that is, li / l '- ^ lvul · -1,

M where v, = [vi, / V2 _; ... vnt.J⁷· Nt self-vectors are also called guide vectors and can be used for self-guidance by the transmitter and for correlated filtering by the receiver, as described below.

The matrix R_m/_So is of dimensions Nt x Nry range N_m¡_Sor, where N_m¡_So &lt;min {Nt, (L + 1)} · The diagonal matrix A_m¡_So therefore contains N_m¡_Sor positive real values ​​along the diagonal, and zeros elsewhere. The largest nonzero input is called the principal self-value A_mi_S0 of matrix R_mys₀ y is indicative of the power gain for the space channel (or "time domain auto-modality") corresponding to that auto-value. The guide vector v_m¡_So independent of frequency, to be used for self-guidance, is the "main" auto-vector of R_myso, which is the column of V_mSo that corresponds to the principal eigenvalue of R_m/_Sor·

The transmitter performs self-guidance on the stream s (n) of data symbols with the guiding vector v_m¡_Sor to obtain Nt flows x_m¡So (n) of transmit symbols, as follows:

. ^{EC (5)}

With the self-guiding shown in equation (5), the data symbol stream s (n) observes an effective channel that is a single input single output (SISO) channel with an effective channel impulse response of h (n) v_m¡_Sor, for n = 0, 1, ..., L. The Nt streams Xmiso (n) of transmission symbols are further processed and transmitted from the Nt transmitting antennas to the receiver.

The receiver gets a flow and_m¡So (n) of symbols received from the only receiving antenna, which can be expressed as:

3Uo («) = á («) ®C «to (ⁿ) ^{+ n}-to (^B)

Eq. (6) + «_et(n) where ® indicates a convolution and n_mis₀(n) is Additive White Gaussian Noise (AWGN). The flow and_mThe or (n) of received symbols experience inter-symbol interference (ISI), which is a phenomenon whereby each symbol in a received stream acts as a distortion for subsequent symbols in the received stream. Inter-symbol interference can be mitigated by using OFDM in conjunction with a sufficiently long cyclic prefix, as described below. Alternatively, for a single carrier MISO system, inter-symbol interference can be mitigated by the use of suitable time-correlated filtering, in combination with equalization, as also described below.

The receiver can perform correlated filtering of the flow and_m¡So (n) of received symbols, either in the time domain or in the frequency domain. The correlated time domain filtering can be expressed as:

Eq. (7)

ES 2 525 141 T3 where ^Sm / so (ⁿ) indicates a detected symbol stream, which is an estimate of the data symbol stream s (n) sent by the transmitter. The correlated filter for n = 0, 1, ..., L, maximizes the received SNR.

An equalizer can be used to mitigate inter-symbol interference due to time spread on the MIMO channel. The equalizer may be a Least Mean Square Error Equalizer (MMSE), Decision Feedback Equalizer (DFE), a Maximum Probability Sequence Estimator (MLSE), or some other type of equalizer. The equalizer can be implemented with an adaptive filter that has coefficients that can be updated with pilot and / or data symbols, and based on specific criteria (eg, least mean square error).

__ Λ

EQ performs EQ on flow ^Sm / so (ⁿ) of detected symbols and provides a stream ^Sm / so (ⁿ) d_andrecovered symbols, which is a best estimate of the data symbol stream s (n) sent by the transmitter. In __ Λ general, the flow^Sm / so (ⁿ) of detected symbols can be provided directly as the stream ^Sm / so (ⁿ) of Λ recovered symbols, or it can be post-processed (e.g. equalized) to obtain the stream ^Sm / so (ⁿ) of recovered symbols.

The correlated frequency domain filtering can be expressed as:

, for k = 1,2 ..... N_F,

Eq. (8) where

Λ ^Sm / so (^) _that| sub-stream of symbols recovered for sub-band k; and ymiso (k) is the sub-stream of received symbols for sub-band k.

The Nf sub-flows and_mSo (k) of received symbols, for k = 1, 2, ..., Λ / f, can be obtained by performing a fast Fourier transformation (FFT) of each set of Nf symbols in the stream and_m¡So (n) of received symbols. The __v«Correlated filter ¹ '* for k = 1, 2, ..., Nf, is a complex-valued scalar that maximizes the

SNR received for each sub-band. The Λ / f symbol sub-streams recovered for the Nf sub-bands can be

Λ multiplexed with each other to obtain the flow ^Sm / so (ⁿ) d_and recovered symbols.

For both time domain and frequency domain correlated filters, the received SNR can be expressed as:

pp Eq. (9)

CXTD _¿total -.H ij „_ ^rtotal to ^οινλΙβ &amp; ο ---- * σ σ where P_totai is the total transmit power used by the transmitter for the data symbol stream; σ² is the lower limit of noise at the receiver; Y

Amiso is the main self-value of R_m/_So c^fl

The frequency independent self-guiding MISO channel capacity can be determined using frequency domain analysis and assuming that the same guide vector is used for all Nf subbands.

c^fl

Capacity can be expressed as:

Giriro - SNR (t)) tal

Eq. (10)

ES 2 525 141 T3 where p is the received mean SNR, as measured at the receive antenna, which is equal to the total receive power divided by the noise σ² of the receiver. The matrix R (k) is the correlation matrix of h (k), which can be obtained and decomposed as follows:

RW = h (*) h (A) = UWA (*) y (it), for k = 1,2 ..... N_F, Eq. (11) where A (k) is the diagonal matrix of self-values ​​of R (k), and

U (k) is the unit matrix of self-vectors of R (k).

The quadratic term in equation (10) can be expressed as:

yLR (*) ^_or= yLu (* ^,

Eq. (12) where z (k) = \ J^H(k '/ v_{m! S}or · Since R (k) has only a non-zero eigenvalue, equation (12) can be simplified as follows:

Lbwí ^ = | z&gt; (*) I'a *), ^{Ec, 13)} where Z (k) is the nonzero eigenvalue of R (k), which is Z (k) = || h (k) ||² for a MISO channel, and zi (k) is the element of z (k) corresponding to the self-value X (k). The MISO channel capacity with frequency independent self-guidance can then be expressed as:

Jtnl

Eq. (14)

C

The capacity ^mt* ° for the MISO channel without self-guidance in the transmitter (or, equivalently, with a guidance vector g = ffÑr v = [gg ... g], where), can be expressed as:

4-1 ** Γ

Eq. (15)

In general, the MISO channel capacity with frequency independent self-guidance is greater than the

C MISO channel capacity without self-guidance.

An exemplary procedure for obtaining the guide vector v_m¡_Sor independent of frequency, for self-guidance at the transmitter in the MISO system. The guide vector can be obtained in other ways as well, and this is within the scope of the invention.

Frequency independent self-guiding can also be used for a MISO system employing OFDM (ie, a MISO-OFDM system). The transmitter can perform time domain self-guiding, as shown in equation (5), where s (n) indicates a sequence of time domain segments for OFDM symbols generated for the modulation data stream. OFDM. OFDM modulation is described later. The transmitter can also perform frequency domain self-guiding on the data symbols for each sub-band, prior to OFDM modulation, to generate OFDM symbols. The receiver can perform correlated filtering in the time domain, as shown in equation (7), or in the frequency domain, as shown in equation (8).

two. MIMO system

A time-dispersing MIMO channel with Nt transmitting antennas and Nr receiving antennas can be characterized by a three-dimensional time-domain channel impulse response matrix H, with dimensions Λ / r χ Λ / τ X (L + 1 ). Matrix H is composed of L + 1 channel impulse response matrices H (n),

ES 2 525 141 T3 for η = 0, 1, L, that is, H = [H (0) H (1) which can be expressed as:

<td> ' £()'</td><td></td><td>Sj («)</td><td>i and w</td><td>Λ</td><td>A *, 00</td>

<td> £()</td><td></td><td>M)</td><td> ^*24 00</td><td>Λ</td><td>Αλ_γ00</td>

<td>M</td><td></td><td>M</td><td>M</td><td>OR</td><td>M</td>

<td></td><td></td><td></td><td></td><td>TO</td><td></td>

for n = 0, 1, L, where the input h¡j (n), for 7 = 1,2, Nr, j = 1,2, ..., Λ /? and n = 0, 1, L, is the coupling (that is, the complex gain) between the transmitting antenna j and the receiving antenna i for time delay n. The row vector h¡ (n), for 7 = 1,2, Λ / ryn = 0, 1, L, includes Nt complex values ​​for the channel gains between the transmitting and receiving antenna / antenna for the time delay n.

The channel impulse response can be estimated by the receiver based on pilot symbols sent by the transmitter. In one embodiment, the transmitter covers the pilot symbol for each transmitting antenna with an orthogonal code assigned to that antenna. The pilot symbols sent from the Nt transmitting antennas are covered by Nt orthogonal codes and can be recovered individually. At the receiver, the pilot symbol received from each receiving antenna i is discovered with the orthogonal codes Nt in a specific time delay, to obtain the channel response between the receiving antenna / 'and each of the Nt transmitting antennas for that time delay, that is, a row of the matrix H (n). The discovery is performed separately for all the Λ / r receiving antennas, to obtain the Λ / r rows of the matrix H (n). The discovery is also performed on L + 1 time delays (that is, for n = 0, 1, L) for each pair of transmitting and receiving antennas, to obtain the L + 1 time domain values ​​for the impulse response. channel for that pair of transmitting and receiving antennas.

A time-dispersive MIMO channel can also be characterized by a corresponding three-dimensional frequency response matrix H of the frequency domain channel, with dimensions Λ / rx Nt * Nf, where Nf&gt; L. The matrix H is composed of Nf matrices H (k) of channel frequency response, for k = 1, 2, Nf, which can be obtained by calculating a discrete Fourier transformation of Nf points on the L + 1 matrices H (n ) channel impulse response, for n = 0, 1, L, as follows:

H (*) =

M *)

M = Σ2 £ («) ·

-o, for k = 1,2 ..... Nf. Eq. (17)

Each matrix H (k), for k = 1, 2, Nf, includes Λ / r row vectors h¡ (k), for / '= 1, 2, Λ / r, for the Λ / r receiving antennas. Each row vector h¡ (k) includes Nt complex values ​​for the channel gains between the Nt transmitting antennas and the receiving antenna i for sub-band k. Each entry h¡j of the matrix H, for / '= 1, 2, Λ / ryj = 1, 2, Nt, includes Nf values ​​of the frequency domain, which can be obtained by taking the DFT of the L + 1 values of the time domain for a corresponding input hj of matrix H. For each pair of transmitting and receiving antennas, there is thus a one-to-one correspondence between the channel impulse response hj and the response hy of channel frequency.

Frequency independent self-guiding can be accomplished in various ways for a MIMO system. Some exemplary frequency independent self-guiding schemes are described below.

A. Self-guide of main mode

For main mode autoguidance, a stream of data symbols is transmitted on the main spatial channel of the MIMO channel, using a single frequency independent guide vector Vpm. To obtain this guide vector, we first compute a matrix R_mm₀dimension correlation Λ / τχ Ντ, as follows:

ES 2 525 141 T3

B = O - * V p joj = w) m * 0 / ·! ** p 4 * 4 ^ 4

Eq. (18)

The decomposition into eigenvalues ​​of R_mmo is then done as follows:

<img file="ES2525141T3_D0002.tif" />

Eq. (19) where V_m/_mo is a unit matrix of eigenvectors of R_mmo and A_m¡Mo is a diagonal matrix whose diagonal entries are the eigenvalues ​​of R_mmine

A MIMO channel can be decomposed into Ns spatial channels, where Ns &lt;min (Nt, Nr}. The matrix R_m/_mo is of rank Ns and the diagonal matrix A_m¡Mo contains Λ / s nonnegative real values ​​along the diagonal. The largest nonzero diagonal entry is called the principal eigenvalue X_m¡Mo of the matrix R_mmine In one embodiment, the guide vector v_p.m to use for self-guidance is the main self-vector of R_m/_mo, which is the column of V_mMo which corresponds to the principal eigenvalue of R_m/_mo. The guide vector v_p.m it can be seen as corresponding to the main spatial channel of the "averaged" MIMO channel.

The transmitter performs self-guidance on the stream s (n) of data symbols with the guide vector Vp_m to get Ntflows x_Pm (ri) of transmit symbols, as follows:

Eq. (20)

With the self-guiding shown in equation (20), the data symbol stream s (n) observes an effective channel that is a single input multiple output (SIMO) channel with an effective channel impulse response of H (n) Vp_m, for n = 0, 1, L. The Nt flows x_p.m(n) of transmitting symbols are further processed and sent from the Nt transmitting antennas to the receiver.

The receiver obtains Λ / r flows j / p_m(n) of symbols received from the Λ / r reception antennas, which can be expressed as:

^{= +} &gt; ^{EC (21)} where_mMo is additive white Gaussian noise, with a mean vector of 0 and a covariance matrix of A_n = or²l · where 0 is a vector of all zeros. The receiver can perform correlated filtering of the j / p streams_m(n) of received symbols, either in the time domain or in the frequency domain.

The correlated time domain filtering can be expressed as:

(n), for n = 0.1 ..... L. Eq. (22)_m í_n\ - v^H and f^to(l, - n)

The correlated filter of the receiver is ~~^pa- 'for n = 0, 1, L, which includes Nr individual correlated filters for the Nr receiving antennas. The correlated filter m_p.m, ¡(N) for each receiving antenna has an impulse response of '-pa-i \ J * p_{ara n} _ q, 1, which maximizes the received SNR for that receiving antenna. The outputs of the Λ / r individual correlated filters for the Nr receiving antennas are summed to obtain the flux^sp.m(ⁿ) of detected symbols. Post-processing (eg, __ Λ equalization) can be performed on the flow^sp.m(ⁿ) of detected symbols to get the stream ^sp.m^l'ⁿ'¹ of symbols retrieved.

The correlated frequency domain filtering can be expressed as:

ES 2 525 141 T3 v (*), for k = 1.2 Nf, Eq. (23) where ypm (k) are the sub-streams of symbols received for sub-band k, which can be obtained by performing an FFT of each set of Nf symbols in the flows and_p.m(n) of received symbols. The correlated filter of the receiver is reception. The filter for k = 1, 2, ..., Nf, which includes Λ / r individual correlated filters for the Λ / r correlated antennas m_pmi¡(K) for each receiving antenna i has a response of for k = 1, 2, ..., Nf. The outputs of the individual correlated Λ / r filters for the Λ / r

Λ receiving antennas for each sub-band A: they are added to obtain the sub-stream ^spnR) ¿θ _sF_mbowls recovered for that sub-band. The Nf sub-streams of symbols recovered for the Nf sub-bands can be multiplexed Λ to obtain the stream^sp.m(ⁿ) d_and recovered symbols.

For both time domain and frequency domain correlated filtering, the received SNR, averaged over the Λ / r receiving antennas, can be expressed as:

Eq. (24)

The correlated filter

<img file="ES2525141T3_D0003.tif" />

, for n = 0, 1, ..., L, it maximizes the received SNR.

The MIMO channel capacity with main mode autoguide can be expressed as:

('raimo you

Eq. (25)

The quadratic term in equation (25) can be expressed as:

Η '(*) Η (*) ν „, = ^ || ν ^ Β, (*) ||^!Α (ί)

1-1

Eq. (26) where uJ / í) is the eigenvector associated with the λ-th eigenvalue λ_λ(/ τ) of the correlation matrix R (/ r) = H^h(/ í) H (/ í). The capacity can then be expressed as:

Eq. (27)

B. Multi-modality self-guide

For multi-modality self-guiding, multiple streams of data symbols are transmitted over multiple spatial channels of the MIMO channel, using multiple frequency-independent guide vectors in a V matrix._mm, whose columns are the eigenvectors of the correlation matrix Rm / mo- Since R_mmo is of rank Ns, where Ns &lt;min {Nr, Nr}, the matrix V_mm can include up to Ns auto-vectors ν_λ, for λ = 1, 2, ..., Ns, for up to Ns autodalities of Rm / mo. For clarity, the following description assumes that all Ns auto-modes are used for data transmission.

The transmitter performs self-guidance on Λ / s flows s_mm(n) of data symbols with guide matrix V_mm, to obtain Nr streams Xmm (n) of transmission symbols, according to the following:

<img file="ES2525141T3_D0004.tif" />

Eq. (28) where Sfr,_m(n) = [si (n) S2 (n) ... SNs (n)] T, V_mm = [vi V2 ... Vns] and Ns &lt;min (Ντ, Nr} for a MIMO channel of range

ES 2 525 141 T3 complete. Each flow s_x(n), for λ = 1,2, Ns, is guided with a respective guide vector v, in the matrix V_mm. Each flow s_x(n) of data symbols observes an effective channel that is a SIMO channel with an effective channel impulse response of Η (η ') ν ,, for n = 0, 1, L. The Λ / rf lux Xmm (n) of transmit symbols are further processed and sent from the Nr transmit antennas to the receiver.

The receiver obtains Λ / r flows and_mm(n) of symbols received from the Λ / r receiving antennas, which and_mm(n) = H (n) ®x_mm(n) + n_m¡_mo(n). For a time-dispersive MIMO channel, self-guiding with multiple frequency-independent guide vectors in the V matrix_mm does not diagonalize the channel. Thus, when multiple spatial channels are used for data transmission using frequency independent auto-guidance, in general, there will be crosstalk between the multiple streams of symbols, as well as inter-symbol interference at the receiver.

The receiver can perform correlated filtering of the flows j /_mm(n) of received symbols, either in the time domain or in the frequency domain. The correlated time domain filtering can be expressed as:

Eq. (29) where -mmW indicates Ns detected symbol streams. The correlated filter of the receiver isⁿ) for n = g, 1, which includes Nr individual correlated filters for the Nr receiving antennas. The correlated filter mmnv (n) for each receiving antenna has an impulse response of 2 «n», í (ⁿ) Xomai ⁿ), for n = o, 1, L. The output of the correlated filter for each receiving antenna includes Ns sub-streams of filtered symbols corresponding to the Ns guide vectors (i.e., Ns columns of V_mm). The Nr sub-streams of filtered symbols from the Nr correlated filters for each guide vector are combined to obtain the stream^s4ⁿ) of symbols detected for that guide vector. Ns detected symbol streams are obtained for the Ns symbols s_mm(n) of data sent by the transmitter.

The correlated frequency domain filtering can be expressed as:

<img file="ES2525141T3_D0005.tif" />

Eq. (30) where and_mm (k) is the sub-stream of received symbols for sub-band k, which is obtained by performing an FFT of each set of Nf symbols in the stream and_mm(n) of symbols received. The correlated filter of the receiver is

M (k} = V^you B^H(k} —j —na— \ i 'p_ara k = 1, 2, ..., Nf, which includes Nr individual correlated filters for the Nr receiving antennas. The correlated filter m_mnv(k) for each receiving antenna has a response of SSniuW for k = 1, 2, ..., Nf. For each sub-band k, the correlated filter output for each receive antenna includes Ns sub-streams of filtered symbols corresponding to the Ns transmit guide vectors. For each sub-band k, the Nr filtered symbol sub-streams from the Nr correlated filters for each guide vector are combined to obtain the detected symbol sub-stream for that guide vector. The Nf sub-streams of symbols detected for the Nf sub-bands for each guide vector are then multiplexed to obtain the stream^s4ⁿ) of symbols detected for stream s_x(n) of data symbols sent with that guide vector. Ns detected symbol streams are obtained for Ns streams s_mm(n) of data symbols sent by the transmitter.

As noted above, if multiple data symbol streams are transmitted simultaneously, then there is crosstalk between these data symbol streams at the receiver for a time-dispersive MIMO channel. A space time equalizer, or "set", can be used to mitigate crosstalk and inter-symbol interference due to time dispersion on the MIMO channel. The space-time equalizer can be a linear least mean squared error (MMSE-LB) equalizer, a decision feedback equalizer (DFE), a maximum probability sequence estimator (MLSE), or some other type of equalizer. . The space and time equalizer can be designed to operate on the Ns symbol streams detected in

Λ both time and space domains, to obtain Ns recovered symbol streams, which are improved estimates of the streams s_mm(n) of data symbols sent by the transmitter. Exemplary designs of the MMSE-LE, the DFE, and the MLSE are described in the commonly assigned U.S. Patent Application No.^s Serial 09 / 993.087, entitled "Multiple Input Multiple Output (MIMO) Access Communication System

ES 2 525 141 T3 multiple ”, filed November 6, 2001.

The space-time equalizer may also implement a receiver processing, successive equalization and interference cancellation technique, which successively recovers one stream of data symbols at a time. As each data symbol stream is recovered, the interference it produces to the remaining data symbol streams, not yet recovered, is estimated and canceled in the detected symbol streams, to obtain "modified" symbol streams. The modified symbol streams are then processed to retrieve the next data symbol stream. The process is repeated until all Ns data symbol streams are recovered. By eliminating interference due to each recovered data symbol stream, data symbol streams not yet recovered experience less interference and can achieve higher SNR values. The receiver processing, successive equalization and interference cancellation technique is also described in the aforementioned US Patent Application with N^s No. 09 / 993.087.

The MIME channel capability with multi-mode autoguide can be expressed as:

<img file="ES2525141T3_D0006.tif" />

^NS

Eq. (31) where | M | indicates the determinant of M. A correlation matrix can be defined as Xeun — WXmn | __{you au}t_or-d values_and can be calculated and indicated as for λ = 1, 2,

Nf. The capacity "^CAT of the MIMO channel with multi-mode autoguide can then be expressed as

<img file="ES2525141T3_D0007.tif" />

Eq. (32)

C. Self-guidance of the main path

For main path self-guiding, a stream of data symbols is transmitted on the main spatial channel for the main propagation path of the MIMO channel, using a single vector v_mp frequency independent guide. As indicated above, a time-dispersing MIMO channel can be characterized by L + 1 channel impulse response matrices H (ri), for n = 0, 1, L. In one embodiment, the main path is defined as the propagation path with the highest received energy. The energy E (n) of each channel impulse response matrix H (n), for n = 0, 1, L, can be calculated as follows:

Α&gt; 0 = Β / £ (») Γ = ££ | \, () Ι '

M, for n = 0.1 ..... L.

Eq. (33)

The energy E (n) is also (1) the trace of the correlation matrix R (n) = tP (n) H (n) and (2) the square of the Frobenius norm of the matrix H (n) of channel impulse response. The highest energy, E_max, for all L + 1 time delays it is then determined as:

Eq. (34)

The delay n_mp of the main path is equal to the time delay of the channel impulse response matrix with the highest energy E_max. The matrix H_mp channel response for the main path is then H_mp = H (n_mp).

A correlation matrix R_mp of H_mp is calculated as correlation matrix R_mp can be expressed as:

<img file="ES2525141T3_D0008.tif" />

The decomposition into eigenvalues ​​of the

R__r = Υ ^ Λ ^ νζ. , ~~^JV~^ap Eq. (₃₅) where V_mp is a unit matrix of eigenvectors of R_mp already_mp is a diagonal matrix of eigenvalues ​​of R_mp.

ES 2 525 141 T3

The matrix R_mpis of rank Ns and the diagonal matrix A_mp contains Λ / s non-negative real values ​​along the diagonal. The vector v_mp frequency independent guide, to be used for self-guidance, is the main self-vector of R_mp, which is the column of V_mp which corresponds to the highest self-value of R_mp.

The transmitter performs self-guidance on the stream s (n) of data symbols with the vector v_mp guide, to obtain Nt streams Xmp (n) of transmission symbols, as follows:

«_(»)=4) ·

Eq. (36)

The spatial processing shown in equation (36) guides the transmit power in the direction of the main spatial channel for the strongest propagation path.

The receiver can perform correlated filtering of streams and_mp (n) of received symbols, which is s and_mp (n) = H (n) ®x_mp(n) + n_mi_mo(n), either in the time domain or in the frequency domain. The correlated time domain filtering can be expressed as:

^ (η) = νζ ^ (Ζ-η) ®ν (n), for n = 0.1 ..... L. Eq. (37) - Λ

The flow ^YeP 'ⁿ'¹ of detected symbols can be post-processed (e.g. equalized) to obtain the flow ^smp (ⁿ) d_andrecovered symbols.

The correlated frequency domain filtering can be expressed as:

, for / (= 1,2 ..... N_F. Eq. (38) s (k \

The Np sub-streams' of recovered symbols, for k = 1, 2, Nf, for the Nf sub-bands, can be Λ multiplexed to obtain the stream ^smpA) d_and recovered symbols.

In general, receiver processing for main path self-guidance can be performed in a similar manner to that described above for main mode self-guidance. However, the correlated filtering is performed based on the vector v_mp guideline for the main spatial channel of the main path, instead of the vector Vp_m guide for the main spatial channel of the "averaged" MIMO channel.

D. Receiver self-guidance

For receiver self-guidance, the MIMO channel is seen as composed of Λ / r MISO channels for Λ / r receiving antennas. Λ / r frequency independent guide vectors can be obtained for the Λ / r channels of MISO, in a similar way to that described above for the MISO system.

As shown in equation (16), the matrix H (n) for the MIMO channel is composed of Λ / r channel impulse response vectors h¡ (n), for i = 1,2, Λ / r . Each row vector h¡ (n) includes the channel impulse response between the Nt transmitting antennas and the receiving / antenna. A correlation matrix R, of dimensions NtX Nt, can be calculated for each receiving antenna, as follows:

<img file="ES2525141T3_D0009.tif" />

Eq. (39)

The decomposition in self-values ​​of the correlation matrix R for each receiving antenna can be carried out as follows:

, for 7 = 1,2 ..... N_R,

Eq. (40)

ES 2 525 141 T3 where

Vj is a unitary matrix whose columns are the eigenvectors of R ,; Y

Λ / is a diagonal matrix whose diagonal entries are the eigenvalues ​​of R /.

Since each h¡ (n) is a row vector for a receiving antenna, the correlation matrix R has rank less than or equal to min {(/. + 1), Ni}. For each receiving / antenna, the vector v_Kií The independent guideline of the frequency that maximizes the received SNR for that receiving antenna is the column of V, which corresponds to the maximum non-zero self-value of R /. Λ / r vectors v_Ki/ guide, for / '= 1,2, Λ / r, are obtained for the Λ / r receiving antennas, and can be represented by a matrix V_K= [v_Kjiv_Kj2 ... v_Kjnr] of dimensions ΝτχΝρ.

One or more data symbol streams can be transmitted with the receiver's guidance. If a stream s (n) of data symbols is transmitted, then the transmitter performs self-guidance on this stream of data symbols, with each of the Λ / r guide vectors, to obtain Ntflows x_to(n) of transmission symbols, as follows:

¡C.

M Eq. (41)

Again, if a stream of data symbols is transmitted, then the receiver can perform correlated filtering of the Λ / r streams and_to(n) of received symbols, which is y_to(n) = H (n) ®x_to(n) + n_m¡_mor (n), either in the time domain or in the frequency domain. For the time domain technique, correlated filtering is performed first for each receiving antenna, as follows:

4m («) ® £ _ &lt;^η), for 7 = 1,2 ..... N_fy,

Eq. (42) where ^srx /ⁿ^ is the filtered symbol stream for the receiving antenna i. The filtered Λ / r symbol streams for all the receiving Λ / r antennas are then combined to obtain the detected symbol stream, as follows:

Eq. (43) ínj

The flow ^s«(ⁿ) of detected symbols can be post-processed (e.g. equalized) to obtain the flow ^srx 'ⁿ'i of recovered symbols, which is an estimate of the transmitted data symbol stream s (n).

For the frequency domain technique, correlated filtering is performed first for each sub-band of each receiving antenna, as follows:

, for 7 = 1,2 ..... N_fyyk = 1,2 ..... N_F,, Eq. (44)

Λ where XA) is the sub-stream of filtered symbols for the sub-band k of the receiving antenna / '. The Λ / r filtered symbol sub-streams for all Λ / r receiving antennas for sub-band k are then combined to obtain the Λ sub-stream^srxW d_and symbols detected for sub-band k, as follows:

<img file="ES2525141T3_D0010.tif" />

, for k = 1,2 ..... N_F.

Eq. (45) s (/ or

The N_F sub-flows ¹ symbols detected for all N_F sub-bands can be multiplexed with each other to Λ obtain the flow ^srx 'ⁿ'i of recovered symbols.

If multiple (Λ / ο) data symbol streams are transmitted, where Ns&gt; Nd&gt; 1, then each data symbol stream can be guided towards a respective set of one or more receiving antennas. The transmitter performs self-guidance for each stream s¿ (n) of data symbols with a set of guide vectors, for a set of antennas

ES 2 525 141 T3 reception, towards which this flow of data symbols is guided, where Ν_λ &gt; 1. The self-guide in the transmitter for each flow Ξ_λ(η) of data symbols can be expressed as:

, for λ = 1,2 ..... Nd,

Eq. (46) where

Xrx, x (n) are the Ntsub-streams of transmit symbols for the stream s¿ (n) of data symbols, and

Vrxxj, for j = 1 ... N¿, are the N¿ guide vectors for the flow s¿ (n) of data symbols.

The Nd sets of Nt transmit symbol sub-streams for all the Nd data symbol streams are then combined to obtain the Nt streams x_rx (n) of transmit symbols, as follows:

<img file="ES2525141T3_D0011.tif" />

The receiver can perform correlated filtering of the flux y, (n) of received symbols for each receiving antenna, either in the time domain, as shown in equation (42), or in the frequency domain, depending on is shown in equation (44). The receiver can then combine the sub-streams^S'^í-J 'ⁿ'i of filtered symbols, for j = 1 ... Νχ, from all receiving antennas used for each stream s¿ (n) of data symbols, to obtain the stream of symbols detected for that stream of data symbols . A space and time equalizer can be used to equalize the Nd streams - ™ (ⁿ) ¿Θ _sF_mboluses detected, to obtain Nd flows -Aⁿ) of recovered symbols.

Frequency independent self-guiding can also be used for a MIMO-OFDM system. The transmitter can perform time domain self-guidance, as shown in equations (20), (28), (36) and (41), where s (h) and s (n) indicate one or more sequences of time domain segments for OFDM symbols generated for the OFDM modulation data stream (s). The transmitter can also perform frequency domain self-guidance on the data symbols for each sub-band, prior to OFDM modulation, to generate OFDM symbols. The receiver can perform correlated filtering in the time domain, as shown in equations (22), (29), (37), (42) and (43). The receiver can also perform frequency domain correlated filtering, as shown in equations (23), (30), (38), (44) and (45).

3. MISO system

FIG. 1 shows a block diagram of a transmitter 110 and a receiver 150 in a MISO system 100. At transmitter 110, a transmitting (TX) data processor 120 receives a stream d (n) of data from a data source 112, processes (eg, encodes, interleaves, and modulates) the data stream according to to a selected transmission mode and provides a stream s (n) of data symbols. The selected transmission mode may be associated with a specific data rate, a specific coding scheme and a specific modulation scheme to be used for the data stream, which are respectively indicated by the data rate, coding and modulation controls provided. by a controller 140.

A TX spatial processor 130 receives the stream s (n) of data symbols and can perform broadband processing, such as spectral spreading or multi-carrier modulation, as described below. The TX spatial processor 130 further performs vector-based self-guidance and_m¡_Sor frequency independent guide (which is also called a TX guide vector) provided by the controller 140. The TX spatial processor 130 also multiplexes the pilot signals with the data and provides Nt streams c_m¡_SW(n) of transmission segments for the Nt transmission antennas. Processing by TX data processor 120 and TX spatial processor 130 is described in more detail below.

A transmitter unit (TMTR) 132 receives and conditions (eg, converts to analog, frequency increases, filters, and amplifies) the Nt streams of transmit segments to obtain Nt modulated signals. Each modulated signal is then transmitted from a respective transmitting antenna (not shown in FIG. 1), and via the MISO channel, to receiver 150. The MISO channel distorts transmitted signals with a channel impulse response h (n) and further degrades transmitted signals with additive white Gaussian noise and possibly interference from other transmission sources.

ES 2 525 141 T3

At receiver 150, the Nt transmitted signals are received by a single receiving antenna (not shown in FIG. 1), and the received signal is provided to a receiver unit (RCVR) 154. Receiver unit 154 conditions and digitizes the received signal to obtain a sample stream for the transmitted data and pilot signals. The receiver unit 154 provides a stream ym¡so (n) of received symbols (for the data) to a receive spatial processor 160 (RX) and the received pilot symbols (for the pilot signal) to a channel estimator 172. The RX spatial processor 160 performs correlated filtering of received symbol stream ymis (n) with a correlated filter and Λ provides a stream^{Sm / so}'ⁿ'¹ of recovered symbols, which is an estimate of the data symbol stream s (n) sent by transmitter 110. An RX data processor 170 then processes (eg, demodulates, deinterleaves, and decodes) the recovered symbol stream according to the selected transmission mode, to obtain a

Decoded data stream d (n), which is an estimate of the data stream d (n) sent by transmitter 110. RX data processor 170 may further provide the status of each received data packet.

The channel estimator 172 processes the received pilot symbols to obtain the channel gain and SNR estimates for the MISO channel. A matrix computing unit 174 then processes the channel gain estimates to obtain the frequency independent guide vector Vmiso for the TX spatial processor 130 and the correlated filter for the RX spatial processor 160. A transmission mode selector 176 receives the SNR estimates from the channel estimator 172, and the packet status from the RX data processor 170, determines a suitable transmission mode for the data stream and provides the transmission mode. selected to a controller 180.

Controller 180 receives guidance vector Vmiso from computing unit 174 and the selected transmission mode from transmission mode selector 176, and assembles the feedback information for transmitter 110. The feedback information is sent to transmitter 110 and used to adjust the processing of the data stream d (n) sent to receiver 150. For example, the transmitter 110 may use the feedback information to adjust the data rate, coding scheme, modulation scheme, self-guiding, or any combination thereof, for the data stream sent to the receiver 150.

Controllers 140 and 180 direct operation at transmitter 110 and receiver 150, respectively. Memory units 142 and 182 provide storage for program codes and data used by controllers 140 and 180, respectively. Memory units 142 and 182 may be internal to controllers 140 and 180, as shown in FIG. 1, or external to these controllers.

FIG. 2 shows a block diagram of one embodiment of the TX data processor 120 in FIG. 1. Within the TX data processor 120, an encoder 212 receives and encodes the data stream d (n) based on an encoding scheme indicated by the encoding control, and provides code bits. The data stream can carry one or more data packets, and each data packet is usually encoded separately, to obtain an encoded data packet. Encryption increases the reliability of data transmission. The encoding scheme may include cyclic redundancy control (CRC) encoding, convolutional encoding, turbo encoding, block encoding, etc., or a combination thereof. A channel interleaver 214 interleaves the code bits based on an interleaving scheme, which may be indicated by an interleaving control if the interleaving depends on the transmission mode. Interleaving provides time, frequency and / or spatial diversity for the code bits.

A symbol mapping unit 216 maps the interleaved bits based on a modulation scheme indicated by the modulation control, and provides a stream of modulation symbols (or simply, "data symbols"). Unit 216 groups each set of interleaved B bits together to form a B-bit binary value, where B&gt; 1, and further maps each B-bit value to a specific modulation symbol based on the modulation scheme (e.g. e.g. QPSK, M-PSK or M-QAM, where M = 2^B). Each modulation symbol is a complex value in a constellation of signals defined by the modulation scheme.

FIG. 3A shows a block diagram of a TX spatial processor 130a, which is an embodiment of the TX spatial processor 130 in FIG. 1. The TX spatial processor 130a includes a self-guiding unit 330, a pilot signal processor 340, and a multiplexer (MUX) 350.

The self-guiding unit 330 includes Nt multipliers 332a to 332t, a multiplier 332 for each of the Nt transmitting antennas. Each multiplier 332 receives the stream s (n) of data symbols and a respective element Vm¡so, j of the TX guide vector Vmiso, multiplies each data symbol by the element Vm¡so, j and provides a stream of symbols of transmission. Multipliers 332a to 332t perform frequency independent self-guiding, as shown in equation (5).

The TX pilot signal processor 340 includes Nt multipliers 342a to 342t, a multiplier 342 for each of the Nt transmit antennas. Each multiplier 342 receives the pilot symbol and a single orthogonal sequence wj assigned to its transmitting antenna, multiplies the pilot symbol by the orthogonal sequence wj and provides a

ES 2 525 141 T3 covered pilot symbol sequence. The multipliers 342a to 342t generate Nt orthogonal pilot signals for the Nt transmit antennas, which can be used for channel estimation by receiver 150.

Multiplexer 350 includes Nt multiplexers 352a to 352t, one multiplexer 352 for each of the Nt transmit antennas. Each multiplexer 352 receives and multiplexes the transmit symbols, coming from an associated multiplier 332, with the covered pilot symbols coming from an associated multiplier 342, and provides a respective stream of transmit segments cj (n). The pilot signal can be multiplexed with the data using time division multiplexing (TDM), as shown in FIG. 3A, code division multiplexing (CDM), subband multiplexing, or some other multiplexing scheme. In any case, the multiplexers 352a to 352t provide Nt streams cj (n) of transmission segments, for J = 1, 2, ..., Nt, for the Nt transmission antennas.

Transmitter unit 132 includes Nt transmitters 362a to 362t, one transmitter 362 for each of the Nt transmit antennas. Each transmitter 362 receives and conditions a respective stream of transmission segments, to generate a modulated signal, which is then transmitted from an associated antenna 134.

FIG. 3B shows a block diagram of a TX spatial processor 130b, which is another embodiment of the TX spatial processor 130 in FIG. 1. TX spatial processor 130b performs time domain spectral spreading and includes spreader 310, self-guiding unit 330, TX pilot signal processor 340, and multiplexer 350.

Within TX spatial processor 130b, spreader 310 receives and spectrally spreads the data symbol stream s (n) with a sequence of pseudo-random numbers (PN) and provides a spread of data symbol stream. Spreading is especially applicable for a low rate data symbol stream, to spectrally spread the data over the entire system bandwidth. The spreading can be performed in a manner similar to that of a CDMA system, which is well known in the art. Self-guiding is then performed on the spread of data symbols stream (instead of the data symbol stream), as described above for FIG. 3A, to obtain Nt transmission segment streams for the Nt transmission antennas.

FIG. 3C shows a block diagram of a TX spatial processor 130c, which is yet another embodiment of the TX spatial processor 130 in FIG. 1. TX spatial processor 130c performs OFDM modulation and includes OFDM modulator 320, self-guiding unit 330, TX pilot signal processor 340, and multiplexer 350.

Within the TX spatial processor 130c, an OFDM modulator 320 receives and performs OFDM modulation on the stream s (n) of data symbols. OFDM effectively divides the overall system bandwidth into multiple (Nr) orthogonal subbands, which are also usually referred to as frequency tones, bins, and sub-channels. With OFDM, each subband is associated with a respective carrier that can be modulated with data. For each OFDM symbol period, one data or pilot symbol may be transmitted for each subband used for transmission, and a signal value of zero is provided for each unused subband. Within the OFDM modulator 320, an inverse fast Fourier transform (IFFT) unit receives a set of data symbols, or pilot, and zeros for the Nr subbands, for each period of OFDM symbols, transforms the set of data symbols, or pilot, and zeros towards the time domain, using an inverse fast Fourier transform, and provides a transformed symbol containing Nr time domain segments. A cyclic prefix generator then repeats a part of each transformed symbol to obtain an OFDM symbol containing Nr + Ncp segments, where Ncp is the number of repeated segments. The cyclic prefix is ​​used to combat selective frequency fading caused by time spreading in the channel. The OFDM modulator 320 provides a stream of data segments for an OFDM symbol stream.

Self-guiding is then performed on the stream of data segments (rather than the stream of data symbols), as described above for FIG. 3A, to obtain Nt transmission segment streams for the Nt transmission antennas. Alternatively, the data symbol stream can be demultiplexed into Ns data symbol sub-streams, and self-guiding can be performed on each data symbol sub-stream. In this case, the same vector v is used_mThis is a guide for all sub-bands. OFDM modulation can then be performed on the autoguide output for all sub-bands of each transmitting antenna, to obtain the stream of transmitting segments for that transmitting antenna. In general, self-guidance can be performed in both the time domain and the frequency domain.

However, time domain self-guidance may require less multiplication and may therefore be easier to implement.

FIG. 4A shows a block diagram of an RX spatial processor 160a, which is an embodiment of the RX spatial processor 160 in FIG. 1, and can be used in conjunction with the TX spatial processor 130a in FIG. 3A. An antenna 152 receives the Nt signals transmitted from the transmitter 110 and provides the signal

ES 2 525 141 T3 received. The receiver unit 154 conditions, digitizes, and pre-processes the received signal, and provides the flow and_m¡So (n) of received symbols. Pre-processing can include filtering, resampling, sample rate conversion, etc.

Within the RX spatial processor 160a, a correlated filter 410 performs the correlated filtering of the stream and the same (n) of received symbols, with the correlated filter ^Piss'_m &amp; o_ \) as shown in equation (7), and provides the stream of detected symbols. An equalizer 412 then performs the equalization

Λ about the stream of detected symbols and provides the stream ^Sm / so (ⁿ) of recovered symbols. Equalizer 412 may implement an MMSE equalizer, a decision feedback equalizer, a maximum probability sequence estimator, or some other type of equalizer, all of which are known in the art. Equalization attempts to mitigate inter-symbol interference due to frequency selectivity on the MISO channel. Correlated filtering and equalization can be integrated with each other (p. eg, correlated filter 410 may be integrated into equalizer 412).

FIG. 4B shows a block diagram of an RX spatial processor 160b, which is another embodiment of the RX spatial processor 160 in FIG. 1. RX spatial processor 160b performs time domain spectral de-spreading and can be used in conjunction with TX spatial processor 130b in FIG. 3B. Inside the RX spatial processor 160b, the correlated filter 410 performs the flow correlated filtering —n)_(n} ymis (n) of symbols received, with the filter correlated and provides the stream “m / sovu ¿θ symbols detected. A de-spreader 412 then de-spreads the stream of symbols detected with (the complex Λ conjugate of) the PN sequence used by transmitter 110 and provides the stream^Sm / so (ⁿ) of recovered symbols. De-spread can be performed with a rake receiver, similar to that of a CDMA system, which is known in the art.

FIG. 4C shows a block diagram of an RX spatial processor 160c, which is yet another embodiment of the RX spatial processor 160 in FIG. 1. RX spatial processor 160c performs OFDM demodulation and can be used in conjunction with TX spatial processor 130c in FIG. 3C. The RX spatial processor 160c includes an OFDM demodulator 420, Nf correlated filters 430a to 430f for the Nf subbands, and a multiplexer432.

Within the RX spatial processor 160c, the OFDM demodulator 420 performs OFDM demodulation on the same stream (n) of received symbols. OFDM demodulator 420 first removes the cyclic prefix on each received OFDM symbol, to obtain a received transformed symbol. The OFDM demodulator 420 then transforms each received transformed symbol into the frequency domain, using a fast Fourier transform (FFT) to obtain a set of received Nf symbols for the Nf subbands. OFDM demodulator 420 provides Nf sub-streams and_mSo (k) of received symbols, for k = 1, 2, ..., Nf, for the Nf subbands, to the Nf correlated filters 430a to 430f. Each correlated filter 430 performs correlated filtering of its sub-stream and_m¡So (k) of received symbols, with its correlated filter -Xmüoll (ty. Q_{Ue is a esca}|_ar ¿Complejo complex value, and provides a sub-flow ^smisoW d_and symbols detected. The multiplexer 432 receives and multiplexes the Nf sub-streams of

Λ symbols detected from all Nffilters correlated 430a to 430f, and provides the flow ^S™ so (ⁿ) d_andrecovered symbols. Less than Nf subbands can be used for data transmission. In this case, the received symbols for the unused subbands are discarded and correlated filtering is not performed for the unused subbands.

FIG. 5 shows a block diagram of a receiver 150x, which is an embodiment of receiver 150 in FIG. 1. The RX spatial processor 160 performs correlated filtering and other pre-processing on the stream and_mare from

Λ received symbols and provides the flow ^S™ so (ⁿ) ¿Θ symbols retrieved to RX data processor 170.

Within the RX data processor 170, a symbol decorrelation unit 512 demodulates the recovered symbols according to the modulation scheme used for the data stream, as indicated by a demodulation control provided by the controller 180. A deinterleaver 514 Channel then de-interleaves the demodulated data in a manner complementary to the interleaving performed at transmitter 110. If the interleaving is dependent on the transmission mode, then the controller 180 provides deinterleaving control to the channel deinterleaver 514. A decoder 516 then decodes the deinterleaved data in a manner complementary to the encoding performed at transmitter 110, as indicated by a decoding control provided by controller 180. For example, a turbo-decoder, or a Viterbi decoder, can be used for decoder 516 if transmitter 110 performs, respectively, turbo-encoding or encoding.

ES 2 525 141 T3 convolutional. Decoder 516 may also provide the status of each received data packet (eg, indicating whether the packet was received correctly or with errors).

Channel estimator 172 obtains pilot symbols received from receiver unit 154, estimates the MISO channel response and the lower limit of noise at receiver 150x, based on the received pilot symbols, and provides the

Λ Λ channel impulse response estimation and estimation ^σ2 from the lower limit of noise to controller 180. Controller 180 performs various functions relating to self-guiding, correlated filtering, and speed control for data transmission. For example, a matrix computing unit 522 within controller 180 performs calculations to obtain the frequency-independent guide vector Vm, or, for transmitter 110, and the correlated filter, for receiver 150. Unit 522 can also estimate the SNR received from the data stream. A transmission mode selector 524 selects a suitable transmission mode for the data stream d (n) based on the received SNR. The memory unit 182 may store a look-up table (LUT) 526 for all transmission modes supported by the MISO system, and their required SNRs. Controller 180 provides the selected transmission mode for the data stream, TX guide vector, acknowledgments (ACK) and / or negative acknowledgments (NAKs), etc., as feedback information for the transmitter 110.

Four. MIMO system

For a MIMO system, Ns spatial channels are available for data transmission, where Ns &lt;min {Nt, Nr}. One data stream can be transmitted on each spatial channel. Each data stream can be processed independently, according to a transmission mode selected for that data stream.

FIG. 6 shows a block diagram of a transmitter 610 and a receiver 650 in a MIMO system 600. In transmitter 610, a TX data processor 620 receives Nd data streams, where Ns&gt; Nd&gt; 1. The TX data processor 620 encodes, interleaves, and modulates each data stream according to its selected transmission mode, and provides a corresponding stream of data symbols. A TX spatial processor 630 receives Nd data symbol streams from TX data processor 620, performs broadband processing (if any) and self-guiding, based on a set of Nd or Nr vector vectors. TX guide provided by controller 640 multiplexes into pilot signal and provides Nt streams of transmit segments for Nt transmit antennas. Processing by TX data processor 620 and TX spatial processor 630 is described in more detail below. A transmitter unit 632 receives and conditions the Nt transmit segment streams to obtain Nt modulated signals, which are transmitted from Nt transmit antennas (not shown in FIG. 6), and via the MIMO channel, to receiver 650.

At receiver 650, the Nt transmitted signals are received by each of the Nr receive antennas (not shown in FIG. 6), and the Nr signals received from the Nr receive antennas are provided to a receiver unit 654. The receiver unit 654 conditions, digitizes and pre-processes each received signal to obtain a corresponding stream of received symbols. Receiver unit 654 provides Nr received symbol streams to RX spatial processor 660, and received pilot symbols to channel estimator 672. The RX spatial processor 660 performs correlated filtering of the Nr received symbol streams, with Nr correlated filters, and provides Nd recovered symbol streams, which are estimates of the Nd data symbol streams sent by the transmitter 610. A processor 670 of RX data processes then (p. eg, demodulates, de-interleaves, and decodes) each recovered symbol stream, according to its transmission mode, to obtain a decoded data stream, which is an estimate of the data stream sent by transmitter 610. Data processor 670 RX can also provide the status of each received data packet.

Channel estimator 672, matrix computing unit 674, and transmission mode selector 676 perform similar functions as channel estimator 172, matrix computing unit 174, and transmission mode selector 176, respectively, in the FIG. 1, to determine the Nd or Nr TX guide vectors for the transmitter 610, the Nr correlated filters for the receiver 650, and the Nd transmission modes for the Nd data streams. A controller 680 assembles feedback information for transmitter 610, which may include the Nd or Nr TX guide vectors and the Nd transmission modes.

Controllers 640 and 680 direct operation on transmitter 610 and receiver 650, respectively. Memory units 642 and 682 provide storage for program codes and data used, respectively, by controllers 640 and 680. Memory units 642 and 682 may be internal to controllers 640 and 680, as shown in FIG. 6, or external to these controllers.

If Nd = 1, then coding, interleaving and modulation for the single data stream can be performed as shown in FIG. 2. Self-guiding for the single data stream can be performed as shown in FIG. 3A, 3B or 3C, depending on whether or not the spectral spreading or OFDM modulation is performed on the data stream. However, self-guiding is performed with the guide vector Vpm for the main mode, or the vector Vmp for

ES 2 525 141 T3 the main path (instead of the guide vector Vmiso). Receiver correlated filtering can be performed as described below. If Nd&gt; 1, then data processing (eg, encoding, interleaving, and modulation) and self-guiding can be performed as described below.

FIG. 7 shows a block diagram of one embodiment of the TX data processor 620 in FIG. 6. For this embodiment, TX data processor 620 includes one set of encoder 712, channel interleaver 714, and symbol mapping unit 716 for each of the Nd data streams. Each set of encoder, channel interleaver and symbol correlation unit receives and processes a respective stream ά_λ(η) data, similar to that described above for TX data processor 120 in FIG. 2, to obtain a corresponding stream srfn) of data symbols. Encoding, interleaving, and modulation for each data stream are performed based on the encoding, interleaving, and modulation controls provided by controller 640, which are generated based on the transmission mode selected for that data stream.

FIG. 8A shows a block diagram of a TX spatial processor 630a, which is an embodiment of the TX spatial processor 630 in FIG. 6, and can be used for multi-mode self-guidance. For this embodiment, the TX spatial processor 630a includes Nd self-guiding units 830a to 830d for the Nd data streams, a TX pilot signal processor 840, a combiner 850, and a multiplexer 860.

For multi-mode self-guiding, each 830 self-guiding unit receives a respective flow s_TO(n) of data symbols and a respective vector ν_λ independent of the frequency in the matrix Vmm. Each self-guiding unit 830 performs self-guidance over its stream of data symbols, with its guidance vector, as described above for FIG. 3A, and provides a respective set of Ntsub-streams χ_λ(η) of transmit symbols for the Nt transmit antennas. The self-guiding units 830a to 830d provide Nd sets of transmit symbol sub-streams for the Nd data streams. Combiner 850 includes Nt combiners 852a through 852t, one combiner 852 for each of the Nt transmitting antennas. Each combiner 852 receives and combines a respective set of Nd transmit symbol sub-streams from the self-guiding units 830a to 830d, for its transmit antenna, and provides a transmit symbol stream. Combiners 852a to 852t provide Nt streams Xmm (n) of transmit symbols for the Nt transmit antennas. Self-guiding units 830a through 830d and combiner 850 collectively perform the self-guiding shown in equation (28).

For receiver self-guiding, the TX spatial processor 630a would include Nr self-guiding units 830, one for each of the Nr receiving antennas. Each self-guiding unit 830 would receive a respective frequency-independent guidance vector νχ in the matrix Vx If Nd = 1, then the same data symbol stream s (n) is provided to all Nr self-guiding units , and is guided with the Nr guide vectors to obtain Nr sets of Nt transmission symbol sub-streams. Each combiner 852 would then receive and combine a respective set of Nr transmit symbol sub-streams from the Nr autoguide units for its transmit antenna, and provide a transmit symbol stream. If Nd&gt; 1, then each data symbol stream can be provided to one or more of the Nr self-guiding units for said one or more antennas, towards which the data symbol stream is guided. Self-guidance is then performed in a similar way to obtain the Nt flows x_rx (n) of transmit symbols for the Nt transmit antennas.

The TX pilot symbol processor 840 receives and covers the pilot symbol with Nt orthogonal sequences, as described above for FIG. 3A, and provides Nt covered pilot symbol sequences for the Nt transmit antennas. The multiplexer 860 includes Nt multiplexers 862a to 862t, one multiplexer 862 for each of the Nt transmit antennas. Each multiplexer 862 receives and multiplexes the transmit symbols from an associated combiner 852 with the covered pilot symbols from an associated multiplexer 842, and provides a respective stream of transmit segments. The multiplexers 862a to 862t provide Nt flows Cmimo (n) = [ci (n) C2 (n) ...¹cnt (n)]^Tof transmission segments for the transmission Ntantenas.

FIG. 8B shows a block diagram of a TX spatial processor 630b, which is another embodiment of the TX spatial processor 630 in FIG. 6. The TX spatial processor 630b performs spectral spreading in the time domain and includes Nd spreaders 810a to 810d for Nd data streams, Nd self-guiding units 830a to 830d, TX pilot symbol processor 840, the Combiner 850 and Multiplexer 860. Each spreader 830 receives and spectrally spreads a respective data symbol stream srfn) with a PN spreading sequence and provides a corresponding spread of data symbol stream. The same, or different, PN sequences can be used for the Nd data symbol streams. The spreaders 810a through 810d provide Nd spread data symbol streams for the Nd data symbol streams. The self-guiding is then performed on each of the Nd spread data symbol streams (instead of the data symbol streams), similar to that described above for FIGS. 3A and 8A, to obtain Nt transmission segment streams for the Nt transmission antennas.

FIG. 8C shows a block diagram of a TX spatial processor 630c, which is yet another embodiment of the TX spatial processor 630 in FIG. 6. The TX spatial processor 630c performs OFDM modulation and

ES 2 525 141 T3 includes Λ / ο OFDM modulators 820a to 820d for Nd data streams, Nd self-guiding units 830a to 830d, TX pilot symbol processor 840, combiner 850 and multiplexer 860.

Each OFDM modulator 820 performs OFDM modulation on a respective stream s¿ (n) of data symbols, similar to that described above for FIG. 3C, and provides a stream of data segments. The OFDM modulators 820a to 820d provide Nd data segment streams for the Nd data streams. Self-guiding is then performed on each of the Nd data segment streams (instead of the data symbol stream), as described above for FIGS. 3A and 8A, to obtain Nt transmission segment streams for the Nt transmission antennas. Alternatively, self-guiding can be performed in the frequency domain on the data symbol sub-stream for each sub-band. In this case, each self-guiding unit uses the same vector ν_λ guide for all sub-bands.

FIG. 9A shows a block diagram of an RX spatial processor 660a, which can be used for the case where a single data stream is transmitted (ie, Nd = 1). Each of the Λ / r receive antennas 652a to 652r receives the Nt signals transmitted from the transmitter 610 and provides a received signal to an associated receiver unit 654. Each receiver unit 654 conditions, digitizes and pre-processes its received signal and provides a stream y, (n) of received symbols.

The RX spatial processor 660a includes Λ / r correlated filters 910a to 91 Or for the Nr receiving antennas, a combiner 912 and an equalizer 914. Each correlated filter 910 performs the correlated filtering of its flow y¡ (n) m₍(n) = ~ⁿ) of received symbols with a mapped filter and provides a filtered symbol stream. The vector v_m¡_mo equals the vector v_p.m guideline for main mode self-guiding, the vector v_mp guide for self-guidance of the main path or vector v_{K &lt;í} guide for self-guidance of the receiver. For receiver self-guidance, each correlated filter 910 is associated with a vector v_KJ different guideline for your receiving antenna, not shown in FIG. 9A. The vector h¡ (n) is the channel impulse response between the transmitting antennas Nt and the receiving antenna / '. The combiner 912 receives and combines the Nr filtered symbol streams from the correlated filters 910a through 91 Or and provides a stream^s™™(ⁿ) d_and detected symbols.

Equalizer 914 performs equalization on the detected symbol stream and provides the stream ^s™2ⁿ¡Of recovered symbols. Equalizer 914 may implement an MMSE equalizer, a decision feedback equalizer, a maximum probability sequence estimator, or some other type of equalizer.

FIG. 9B shows a block diagram of an RX spatial processor 660b, which can also be used for the case where a single data stream is transmitted (ie, Nd = 1). The RX spatial processor 660b performs spectral de-spreading in the time domain and can be used in conjunction with the TX spatial processor 630b in FIG. 8B.

The RX spatial processor 660b includes Nr correlated filters 910a to 91 Or for the Nr receive antennas, the combiner 912, and a decoder 916. Each correlated filter 910 performs correlated filtering of a respective stream y, (n) of received symbols with its correlated filter (f ' ⁿ) And p_rop_orc¡_{ona a} filtered symbol stream. Combiner 912 receives and combines the Nr filtered symbol streams and provides the stream^smmo (ⁿ) of detected symbols. The de-spreader 916 then de-spreads the detected symbol stream with the

Λ PN sequence used by the 610 transmitter and provides the flow ^s™ no (ⁿ) ¿Θ symbols retrieved.

FIG. 9C shows a block diagram of an RX spatial processor 660c, which can also be used for the case where a single data stream is transmitted (ie, Nd = 1). The RX spatial processor 660c performs OFDM demodulation and can be used in conjunction with the TX spatial processor 630c in FIG. 8C.

The RX spatial processor 660c includes Nr correlated antenna filters 920a to 920r for the Nr receive antennas, Nf combiners 932a to 932f for the Nf subbands, and a multiplexer 934. Each correlated antenna filter 920 performs correlated filtering for a receiving antenna and includes an OFDM demodulator 922 and Nf correlated filters 930a to 930f for the Nf sub-bands.

Within each antenna correlated filter 920, OFDM demodulator 922 performs OFDM demodulation on the stream y, (n) of received symbols for the associated receive antenna and provides Nf sub-streams y, (k) of received symbols. , for k = 1, 2, Nf, for the Nf subbands, a Nf correlated filters 930a to 930f. Each correlated filter 930 performs the correlated filtering of its sub-stream y, (k) of symbols received with its filter

ES 2 525 141 T3 correlated ^— and provides a sub-stream of filtered symbols. The vector h¡ (k) is the channel frequency response for the receiving antenna i for sub-band k. The correlated filters 930a to 930f for each antenna correlated filter 920 provide Nf filtered symbol sub-streams, for the Nf subbands, to Nf combiners 932a to 932f.

Each combiner 932 receives and combines the Nr filtered symbol sub-streams from the Nr antenna correlated filters 920a to 920r for its sub-band, and provides a detected symbol sub-stream for the sub-band. The multiplexer 934 receives and multiplexes the Nf detected symbol sub-streams from the Λ combiners 932a to 932f for the Nf sub-bands and provides the stream d_and recovered symbols.

FIG. 9D shows a block diagram of an RX 660d spatial processor, which can be used for multi-mode auto-guidance with Afo&gt; 1 The RX 660d spatial processor can be used in conjunction with the TX 630a spatial processor in the FIG. 8A, or TX spatial processor 630b in FIG. 8B.

The RX 660d spatial processor includes Nr correlated filters 940a to 940r for the Nr receive antennas, Nd combiners 942a to 942d for the Nd data streams, a space-time equalizer 944, and Nd de-spread 946a to 946d for the Nd streams. data. Each correlated filter 940 performs the filtering w (¿1 = (L — ri \ correlated from a respective stream y¡ (n) of received symbols with a correlated filter “* —í for an associated receiving antenna, and provides Nd sub- filtered symbol streams for the Nd data symbol streams.The matrix V_mm includes Nd vectors v, guide, for λ = 1,2, ..., Afo, for the Nd data symbol streams. Each correlated filter 940 therefore performs the correlated filtering of the flow y, (n) of symbols received with m_the(n) (Ln),

Nd correlated filters, for λ = 1,2, Nd, where ν_λ is the λ-th column of V_mm, to obtain the Nd filtered symbol sub-streams for the associated receive antenna.

Each combiner 942 receives and combines the Nr filtered symbol sub-streams from the correlated filters 940a to 940r for a data symbol stream, and provides the stream Tⁿ'D_and symbols detected for the data stream. The correlated filters 940a through 940r, and the combiners 94-2a through 942d, collectively perform the correlated filtering shown in equation (29), and provide Nd detected symbol streams for the Nd data symbol streams.

If multiple data symbol streams are transmitted, then there is likely crosstalk between these data symbol streams at receiver 650. Space-time equalizer 944 performs equalization on the Nd detected symbol streams from combiners 942a to 942d, and provides Nd equalized symbol streams. The space-time equalizer 944 may implement an MMSE linear equalizer, a decision feedback equalizer, a maximum probability sequence estimator, or some other type of equalizer that can operate together over multiple streams to mitigate crosstalk and / or or maximize received SNR in the presence of crosstalk, inter-symbol interference, and noise. The space time equalizer 944 may also implement the interference cancellation and successive equalization processing technique. The space and time equalizer 944 can also be omitted.

If spreading is not performed at transmitter 610, as shown in FIG. 8A, then the Nd equalized symbol streams from the space-time equalizer 944 are provided as the Nd recovered symbol streams Λ. If spreading is performed at transmitter 610 for each data symbol stream, as shown in FIG. 8B, then each de-spreader 946 receives and de-spreads a respective stream of symbols equalized with the PN sequence, and provides a corresponding stream of recovered symbols. The

* De-spreaders 946a to 946d will then provide the Nd recovered symbol streams.

FIG. 9E shows a block diagram of an RX 660e spatial processor, which can be used for receiver self-guidance with Nd&gt; 1. The RX 660e spatial processor includes Nr correlated filters 950a to 950r for the Nr receive antennas, a combiner 952, a space time equalizer 954 and the Nd de-spreaders 956a to 956d for the Nd data symbol streams. Each correlated filter 950 performs correlated filtering of a respective stream y, (n) of received symbols with a correlated filter for an associated receive antenna, and provides a stream of filtered symbols. Combiner 952 receives the Nr filtered symbol streams from correlated filters 950a to 950r, combines the filtered symbol streams for all receiving antennas used for each data symbol stream, and provides the stream Tⁿ'¡Of symbols detected for that data symbol stream. The combination is

ES 2 525 141 T3 dependent on the self-guidance performed in the transmitter (that is, the specific reception antennas towards which each stream of data symbols is guided). Combiner 952 provides Nd detected symbol streams for the Nd data symbol streams. The space-time equalizer 954 and de-spreaders 956a to 956d operate on the Nd detected symbol streams, as described above for FIG. 9D, and provide the

Λ

Nd retrieved symbol streams.

FIG. 9F shows a block diagram of an RX 660f spatial processor, which can also be used for multi-mode autoguiding with Nd&gt; 1. The RX 660f spatial processor performs OFDM demodulation and can be used in conjunction with TX spatial processor 630c in FIG. 8C.

The RX spatial processor 660f includes the Nr correlated antenna filters 970a to 970r, for the Nr receive antennas, the Nf combiners 982a to 982f for the Nf sub-bands, the Nf space and time equalizers 984a to 984f for the Nf subbands and a multiplexer 986. Each antenna correlated filter 970 performs correlated filtering for a receiving antenna and includes an OFDM demodulator 972 and Nf correlated filters 980a to 980f for the Nf subbands.

Within each antenna correlated filter 970, the OFDM demodulator 972 performs OFDM demodulation on the stream y, (n) of received symbols for the associated antenna, and provides Nf sub-streams y, (k) of received symbols, for k = 1, 2, ..., Nf, for the Nf sub-bands, to the Nf correlated filters 980a to 980f. Each correlated filter 980 performs the correlated filtering of its sub-stream y, (k) of received symbols with its correlated filter —'— o »» -¹ '' and provides Nd filtered symbol sub-streams for the Nd data symbol streams for its sub-band. The correlated filters 980a to 980f for each antenna correlated filter 970 provide Nf sets of Nd filtered symbol sub-streams, for the Nf subbands, to the Nf combiners 982a to 982f.

Each combiner 982 receives and combines the Nr sets of Nd filtered symbol sub-streams from the Nr antenna correlated filters 970a to 970r, for its sub-band, and provides Nd detected symbol sub-streams for its sub-band. Although not shown in FIG. 9F, each combiner 982 includes Nd adders, one adder for each data symbol stream. Each adder receives and sums the Nr filtered symbol sub-streams from the antenna correlated filters 970a to 970r for its sub-band, and its data symbol stream, to obtain the detected symbol sub-stream for its sub-band. band.

Each spatial equalizer 984 performs equalization on the detected Nd symbol sub-streams from an associated combiner 982 for its sub-band, and provides Nd equalized symbol streams for the sub-band. Spatial equalizer 984 may implement an MMSE linear equalizer or some other equalizer that operates in conjunction on multiple symbol streams to mitigate crosstalk and / or maximize received SNR. The spatial equalizer 984 may also implement the interference cancellation and successive equalization processing technique.

Multiplexer 986 receives Nf sets of Nd equalized symbol sub-streams from combiners 984a to 984f for the Nf subbands. The multiplexer 986 then multiplexes the Nf equalized symbol sub-streams from

• Combiners 984a to 984f for each data symbol stream, and provides the retrieved symbol stream for that data symbol stream.

FIG. 10 shows a block diagram of a 650x receiver, which is an embodiment of receiver 650 in FIG. 6. The RX spatial processor 660 performs correlated filtering and post-processing on the Nr streams y, (n) of

Λ 'received symbols, for 7 = 1,2, ..., Nr, and provides Nd streams ^SM) ¿θ symbols recovered, for λ = 1, 2, ..., Nd, to the RX data processor 670. The RX data processor 670 includes a set of symbol de-spreading unit 1012, deinterleaver 1014, and decoder 1016, for each of the Nd recovered symbol streams. Each set of symbol despread, deinterleaver, and decoder unit processes a respective stream of recovered symbols, as described above for FIG. 5. The 670 data processor

Λ of RX provides Nd flows θ d_tot_you decoded, for λ = 1, 2, ..., Afo.

The channel estimator 672 estimates the channel response and the lower limit of receiver noise, based on the symbols

Λ pilot received from receiver units 654a to 654r, and provides the estimate —A) of the response of ^Λ 2 channel boost and estimation ^σ from the lower limit of noise to the controller 680. The controller 680 performs various functions relating to self-guiding, correlated filtering, and speed control for data transmission. For example, a matrix computing unit 1022 can perform calculations to obtain (1) the vector v_p.m guideline for the main mode self-guidance, (2) the guide vector Vmp for the main path self-guidance, (3) the Nd vectors ν_λ

ES 2 525 141 T3 guide, for λ = 1, 2, Nd, for multi-modality self-guide, or (4) the Nr vectors v_Kguide j, for i = 1, 2,

Nr, for the receiver's self-guidance. The calculation unit 1022 also obtains Nr correlated filters for the receiver 650 and can further estimate the received SNR of the Nd data streams. A transmission mode selector 1024 selects a suitable transmission mode for each data stream, based on its received SNR. The memory unit 682 may store a look-up table 1026 for all supported transmission modes, and their required SNRs. The controller 680 provides the Nd TX guide vectors, the Nd transmission modes selected for the Nd data streams, the ACKs and / or the NAKs, etc., as feedback information to the transmitter 610.

For the embodiments described above, the receiver estimates the channel response of the MISO or MIMO channel, obtains the guide vector, or vectors, for the transmitter and the filter, or filters, correlated for the receiver, and returns the guide vector (s) as feedback information. For other embodiments, it may be possible for the transmitter to estimate the channel response and obtain the guide vector (s). For example, in a time division duplexed (TDD) system with a shared frequency band, the downlink and uplink channel responses can be assumed to be reciprocal with each other. That is, if H (k) represents a channel frequency response matrix from antenna array A to antenna array B for sub-band k, then a reciprocal channel implies that the coupling from array B to the formation A is given by H^T(k). For the TDD system, the channel reciprocal characteristics can be exploited to allow the transmitter to estimate the link observed by the receiver, based on the pilot signal sent by the receiver on the other link. In general, the channel estimation and the calculation of the guide vectors can be performed by the receiver or the transmitter, depending on the design of the system.

FIG. 11 shows a flow chart of one embodiment of a process 1100 for performing main mode self-guidance, multi-mode self-guidance, and main path self-guidance in a MIMO system. Initially, a plurality of channel response matrices are obtained for the channel response of the MIMO channel in the MIMO system (at block 1112). These channel response matrices can be (1) L + 1 channel impulse response matrices for L + 1 time delays (i.e., H (n) for n = 0, 1, ..., L) or ( 2) Nf channel frequency response matrices for Nf subbands (that is, H (k) for k = 1,2, ..., Nf).

A single correlation matrix is ​​calculated for the MIMO channel, based on the channel response matrices (at block 1114). For the main modality autoguide and the multi-modality autoguide, the single correlation matrix can be obtained (1) by calculating a correlation matrix of each between the plurality of channel response matrices and (2) adding the correlation matrices for the channel response matrices, to obtain the single correlation matrix, as shown in equation (18). For the self-guiding of the main path, the unique correlation matrix can be obtained (1) determining the energy of each of the channel impulse response matrices, (2) identifying the channel impulse response matrix with the increased energy, (3) calculating a correlation matrix of the channel impulse response matrix with the highest energy and (4) defining the only correlation matrix for the MIMO channel as the correlation matrix of the channel impulse response matrix with the highest energy, as shown in equations (33) and (34) and the related description.

The single correlation matrix is ​​then decomposed (e.g., using self-value decomposition) to obtain Nd guide vectors for Nd spatial channels of the MIMO channel, where Ns&gt; Nd&gt; 1 and Ns is the number of automodalities. of the single correlation matrix (at block 1116). For the self-guidance of the main mode and the self-guidance of the main path, Nd = 1 and only one guidance vector is obtained. For multi-modality self-guidance, Nd&gt; 1 and multiple guide vectors are obtained.

The operations shown in blocks 1112, 1114, and 1116 can be performed by receiver 650 in FIG. 6. The operations shown in blocks 1112, 1114, and 1116 can also be performed by transmitter 610 for a time division duplexed (TDD) system, in which the downlink and uplink share the same frequency band. In either case, the Nd guide vectors can be used for self- guidance by the transmitter and for correlated filtering by the receiver.

At the transmitter, each guide vector can be used for frequency independent self-guidance or spatial processing of a data stream sent on the spatial channel associated with the guide vector (at block 1122). The transmitter performs self-guiding over Nd data symbol streams with the Nd guide vectors, to generate Nt transmit symbol streams (at block 1124), which are further processed and transmitted from the Nt transmit antennas (at block 1124). block 1126).

At the receiver, correlated filtering of Nr received symbol streams for Nr receiving antennas, where Nr&gt; Nd, can be performed in both the time domain and the frequency domain. A correlated filter can be obtained for each receiving antenna, based on the Nd guide vectors and a plurality of channel response vectors for that receiving antenna (at block 1132). The channel response vectors for each receiving antenna can be obtained from the channel response matrices. The flow of symbols

ES 2 525 141 T3 received for each receiving antenna is filtered with the correlated filter for that receiving antenna, to obtain Nd sub-streams of filtered symbols, a sub-stream for each guide vector used by the transmitter (in the block 1134). The filtered symbol sub-streams from all Nr correlated filters for the Nr receive antennas are then combined to obtain Nd detected symbol streams for the Nd data streams sent by the transmitter (at block 1136). Equalization can be performed on the Nd detected symbol streams to obtain Nd recovered symbol streams (at block 1138). If Nd&gt; 1, then the space-time equalization (p. eg with an MMSE-LE, a DFE or an MLSE) can be performed on the multiple detected symbol streams to obtain multiple recovered symbol streams.

FIG. 12 shows a flow diagram of one embodiment of a 1200 process to perform receiver self-guidance in a MISO or MIMO system with Nt transmitting antennas and Nr receiving antennas, where Nt&gt; 1 and Nr&gt; 1 in this case. Initially, Nr sets of channel response vectors are obtained for the Nr receive antennas, one set for each receive antenna (at block 1212). Each set of channel response vectors is indicative of the channel frequency response or channel impulse response between the Nt transmit antennas and a receive antenna.

A single correlation matrix is ​​calculated for each receive antenna, based on the set of channel response vectors for that receive antenna (at block 1214). This can be achieved by (1) calculating a correlation matrix of each of the channel response vectors for the receiving antenna and (2) adding correlation matrices for the channel response vectors for the receiving antenna, to obtain the Unique correlation matrix for the receiving antenna. The single correlation matrix for each receiving antenna is then decomposed (eg, using eigenvalue decomposition) to obtain a guide vector for the receiving antenna (at block 1216). The operations shown in blocks 1212, 1214, and 1216 can be performed by receiver 150 in FIG. 1 or receiver 650 in FIG. 6. The operations shown in blocks 1212, 1214, and 1216 can also be performed by transmitter 110 in FIG. 1 or transmitter 610 in FIG. 6 for a TDD system. In any case, Nr guide vectors are obtained for the Nr receiving antennas, and can be used for spatial processing by the transmitter, and for correlated filtering by the receiver.

At the transmitter, the Nr guide vectors can be used for frequency independent self-guidance or spatial processing of Nd data streams, where Nr&gt; Nd&gt; 1 and Nt&gt; Nd (at block 1222). For a MISO system with only one receiving antenna (Nr = 1), a data stream is sent using a guide vector obtained for said receiving antenna (in blocks 1224 and 1226). For a MIMO system with a plurality of receiving antennas (Nr&gt; 1), one or more data streams can be sent using the Nr guide vectors obtained for the Nr receiving antennas. Each data stream can be guided to one or more receiving antennas.

At the receiver, the correlated filtering of Nr received symbol streams for the Nr receive antennas can be performed in both the time domain and the frequency domain. A correlated filter is obtained for each receiving antenna, based on the guide vector and the set of channel response vectors for that receiving antenna (at block 1232). The received symbol stream for each receiving antenna is filtered with the correlated filter for that receiving antenna, to obtain a filtered symbol stream for the receiving antenna (at block 1234). The Nr filtered symbol streams from the Nr correlated filters for the Nr receive antennas are then combined to obtain Nd detected symbol streams for the Nd data streams sent by the transmitter (at block 1236). Equalization can be performed on the Nd detected symbol streams to obtain Nd recovered symbol streams for the Nd data streams sent by the transmitter (at block 1238).

5. Speed ​​selection

For both MISO 100 and MIMO 600 systems, the receiver can estimate the received SNR for each spatial channel. The SNR calculation may depend on the self-guiding scheme used for data transmission, as described above. The receiver can then calculate an operational SNR, γορ (λ), for spatial channel, based on the received SNR, γΧλ), and an SNR shift, γο₅(λ), for the spatial channel (eg, γο₅(λ) = γΧλ) + γο₅(λ), where the units are in dB). The SNR shift can be used to compensate for estimation error, channel variability, and other factors. The receiver can select a suitable transmission mode for each spatial channel, based on the operational SNR for that spatial channel.

The system can be designed to support a set of transmission modes. One of the supported transmission modes may be for zero rate (ie, zero data rate). Each of the remaining transmission modes is associated with a specific non-zero data rate, a specific coding scheme or code rate, a specific modulation scheme and a specific minimum SNR required to achieve the desired level of performance (p. eg 1% Packet Error Rate (PER)) for an AWGN channel without fading. For each supported transmission mode, with a non-zero data rate, the required SNR is obtained based on the specific design of the system (that is, the specific rate of

ES 2 525 141 T3 code, the interleaving scheme, the modulation scheme, etc., used by the system for that transmission mode) and for an AWGN channel. The required SNR can be obtained by computer simulation, empirical measurements, etc., as is known in the art. The set of supported transmission modes and their required SNRs can be stored in a look-up table.

The operational SNR, γ_ορ(λ), for each spatial channel can be provided to the lookup table, which then provides the mode q (X) of transmission for that spatial channel. This mode q (X) of transmission is the supported transmission mode with the highest data rate and a required SNR, γ ^ (λ), which is less than or equal to the operational SNR (i.e., γ ^ (λ ) &lt;γορ (λ)). The receiver thereby selects the highest possible data rate for each spatial channel, based on the operating SNR for that spatial channel.

For clarity, specific embodiments of various autoguide schemes have been described above. Other variants of these self-guiding schemes can also be devised, and this is within the scope of the invention. For example, the single correlation matrix for the MIMO channel can be calculated in other ways than described above for the main mode and multi-mode self-guiding schemes. As another example, multiple streams of data symbols may be transmitted over multiple spatial channels on the main path. As yet another example, Nd data symbol streams can be transmitted on the Nd best spatial channels, based on the energy of the spatial channels. Other self-guidance schemes may also be devised, based on the teachings provided herein, and this is within the scope of the invention.

The self-guiding techniques described herein can be implemented by various means. For example, these techniques can be implemented in hardware, software, or a combination thereof. For a hardware implementation, processing at the transmitter for self-guiding and other pertinent functions can be implemented within one or more application-specific integrated circuits (ASICs), digital signal processors (DSP), signal processing devices. digital signals (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), processors, controllers, microcontrollers, microprocessors, other electronic units designed to perform the functions described herein, or a combination thereof. Processing at the receiver for correlated filtering and other pertinent functions can also be implemented within one or more ASICs, DSPs, etc.

For a software implementation, self-guiding techniques can be implemented with modules (eg, procedures, functions, etc.) that perform the functions described herein. Software codes can be stored in memory units (eg, memory units 142 and 182 in FIG. 1 or memory units 642 and 682 in FIG. 6) and executed by a processor (eg. eg, controllers 140 and 180 in FIG. 1 or controllers 640 and 680 in FIG. 6). The memory unit may be implemented within the processor or external to the processor, in which case it may be communicatively coupled with the processor by 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 below, and these concepts may have applicability in other sections throughout the entire specification.

The foregoing 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 immediately apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the scope of the invention, as defined by the appended claims. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be accorded the broadest scope consistent with the novel principles and features disclosed herein, as defined by the claims.

Specific embodiments that are not part of the invention

1. A method of performing spatial processing in a multiple input, multiple output (MIMO) communication system, comprising: obtaining a plurality of channel response matrices for a channel response of a MIMO channel in the MIMO system; calculating a correlation matrix for the MIMO channel based on the plurality of channel response matrices; and decomposing the correlation matrix to obtain at least one guide vector for at least one spatial channel of the MIMO channel, wherein said at least one guide vector is used by a transmitting entity for frequency independent spatial processing of a data flow sent by said at least one spatial channel associated with said at least one guide vector.

two. The method of Embodiment 1, wherein the plurality of channel response matrices comprise a plurality of channel impulse response matrices for a plurality of time delays of a response

MIMO channel channel pulse ES 2 525 141 T3.

3. The method of Embodiment 1, wherein the plurality of channel response matrices comprise a plurality of channel response matrices for a channel frequency response for a plurality of subbands of the MIMO channel.

Four. The method of Embodiment 1, in which the calculation of the correlation matrix for the MIMO channel includes: calculating a correlation matrix of each among the plurality of channel response matrices, to obtain a plurality of correlation matrices for the plurality of channel response matrices, and summing the plurality of correlation matrices for the plurality of channel response matrices, to obtain the correlation matrix for the MIMO channel.

5. The method of embodiment 2, in which the calculation of the correlation matrix for the MIMO channel includes: determine the energy of each of the plurality of channel impulse response matrices, identify a channel impulse response matrix with the highest energy among the plurality of channel impulse response matrices, and calculate a correlation matrix of the channel impulse response matrix with the highest energy, to generate the correlation matrix for the MIMO channel.

6. The method of embodiment 1, in which the decomposition into self-values ​​of the correlation matrix is ​​performed to obtain said at least one guide vector for said at least one spatial channel of the MIMO channel.

7. The method of embodiment 1, further comprising: sending said at least one guide vector as feedback information to the transmitting entity.

8. The method of embodiment 1, in which said at least one guide vector is used by the transmitting entity to generate a plurality of streams of transmission segments for at least one data stream sent by said at least one spatial channel of the channel of MIMO, and in which the plurality of streams of transmitting segments are transmitted from a plurality of transmitting antennas at the transmitting entity.

9. The method of Embodiment 1, in which frequency independent spatial processing is performed by the transmitting entity in the time domain on a stream of time domain segments, generated for the OFDM modulation data stream.

10. The method of embodiment 1, in which frequency-independent spatial processing is performed by the transmitting entity in the frequency domain for each of a plurality of subbands, on data symbols generated for the data stream .

eleven. The method of embodiment 1, further comprising: obtaining, from the plurality of channel response matrices, a plurality of channel response vectors, for each of a plurality of receiving antennas at a receiving entity; and obtaining a correlated filter for each of the plurality of receiving antennas, based on said at least one guide vector and the plurality of channel response vectors for the respective receiving antenna.

12. The method of Embodiment 11, in which the correlated filter for each among the plurality of receiving antennas is used to maximize the received signal-to-noise ratio (SNR) for the respective receiving antenna.

13. The method of embodiment 11, further comprising: filtering a plurality of received symbol streams for the plurality of receiving antennas, with the plurality of correlated filters.

14. The method of Embodiment 13, wherein the plurality of channel response matrices comprise a plurality of channel impulse response matrices for a plurality of time delays of a channel impulse response of the MIMO channel, and in the which filtering is performed in the time domain with a plurality of correlated time domain filters, obtained for the plurality of receiving antennas based on said at least one guide vector and the plurality of channel impulse response matrices.

fifteen. The method of embodiment 13, wherein the plurality of channel response matrices comprise a plurality of channel frequency response matrices, for a channel frequency response, for a plurality of MIMO channel subbands, and in which the filtering is performed in the frequency domain with a plurality of correlated frequency domain filters, obtained for the plurality of receiving antennas on the basis of said at least one guide vector and the plurality of channel frequency response matrices.

16. The method of embodiment 1, in which a guide vector is obtained and used by the transmitting entity for frequency-independent spatial processing of a data stream.

17. The method of embodiment 16, further comprising: obtaining a correlated filter for each

ES 2 525 141 T3 one of a plurality of receiving antennas in a receiving entity, based on said guide vector and a plurality of channel response vectors for the receiving antenna, wherein the plurality of channel response vectors for each receiving antenna they are obtained from the plurality of channel response matrices; filtering a plurality of received symbol streams for the plurality of receiving antennas with the plurality of correlated filters, to obtain a plurality of filtered symbol streams; and combining the plurality of filtered symbol streams to obtain a detected symbol stream for said data stream sent by the transmitting entity.

18. The method of embodiment 17, further comprising: performing equalization on the detected symbol stream to obtain a recovered symbol stream for said data stream.

19. The method of embodiment 1, in which a plurality of guide vectors are obtained and used by the transmitting entity for frequency-independent spatial processing of a plurality of data streams sent on a plurality of spatial channels associated with the plurality of guide vectors.

twenty. The method of embodiment 19, further comprising: obtaining a correlated filter for each among a plurality of receiving antennas at a receiving entity, based on the plurality of guide vectors and a plurality of channel response vectors for the receiving antenna, wherein the plurality of channel response vectors for each receiving antenna are obtained from the plurality of channel response matrices; filtering a plurality of received symbol streams for the plurality of receiving antennas with the plurality of correlated filters, to obtain a plurality of filtered symbol sub-streams; and combining the plurality of filtered symbol sub-streams to obtain a plurality of detected symbol streams for the plurality of data streams sent by the transmitting entity.

twenty-one. The method of embodiment 20, further comprising: performing space and time equalization for the plurality of detected symbol streams, to obtain a plurality of recovered symbol streams for the plurality of data streams.

22. The method of embodiment 21, in which the space and time equalization is performed with a linear least mean square error equalizer (MMSE-LE), a decision feedback equalizer (DFE) or a sequence estimator of maximum probability (MLSE).

2. 3. An apparatus in a multi-input, multiple-output (MIMO) wireless communication system, comprising: a channel estimator for obtaining a plurality of channel response matrices for a channel response of a MIMO channel in a MIMO system ; and a controller for calculating a correlation matrix for the MIMO channel, based on the plurality of channel response matrices, and decomposing the correlation matrix to obtain at least one guide vector for at least one spatial channel of the channel. MIME, wherein said at least one guide vector is used by a transmitting entity for frequency-independent spatial processing of a data stream sent by said at least one spatial channel associated with said at least one guide vector.

24. The apparatus of embodiment 23, in which the controller calculates a correlation matrix of each among the plurality of channel response matrices, to obtain a plurality of correlation matrices for the plurality of channel response matrices, and summing the plurality of correlation matrices to obtain the correlation matrix for the MIMO channel.

25. The apparatus of embodiment 23, wherein the plurality of channel response matrices comprise a plurality of channel impulse response matrices for a plurality of time delays of a channel impulse response of the MIMO channel, and in the which the controller determines the energy of each among the plurality of channel impulse response matrices, and calculates a correlation matrix of a channel impulse response matrix with the highest energy among the plurality of channel impulse response matrices to be obtained.

26. The apparatus of embodiment 23, further comprising: a plurality of correlated filters for a plurality of receiving antennas, one correlated filter for each receiving antenna, and each correlated filter is used to filter a stream of received symbols for an associated receiving antenna, to obtain a stream of filtered symbols, wherein the correlated filter for each receiving antenna is obtained based on said at least one guide vector and a plurality of channel response vectors for the receiving antenna, and wherein the plurality of channel response vectors for each receiving antenna are obtained from the plurality of channel response matrices; and a combiner for combining a plurality of filtered symbol streams from the plurality of correlated filters, to obtain at least one detected symbol stream for at least one data stream sent by the transmitting entity.

27. An apparatus in a multiple input, multiple output (MIMO) wireless communication system, comprising: means for obtaining a plurality of channel response matrices for a channel response of a MIMO channel in the MIMO system; means of calculating a correlation matrix for the MIMO channel based on

ES 2 525 141 T3 to the plurality of channel response matrices; and means for decomposing the correlation matrix to obtain at least one guide vector for at least one spatial channel of the MIMO channel, wherein said at least one guide vector is used by a transmitting entity for frequency independent spatial processing. of a data stream sent by said at least one spatial channel associated with said at least one guide vector.

28. The apparatus of embodiment 27, wherein the means for calculating the correlation matrix includes: means for calculating a correlation matrix of each among the plurality of channel response matrices, to obtain a plurality of correlation matrices for the plurality of channel response matrices, and means for summing the plurality of correlation matrices, to obtain the correlation matrix for the MIMO channel.

29. The apparatus of embodiment 27, wherein the plurality of channel response matrices comprise a plurality of channel impulse response matrices for a plurality of time delays of a channel impulse response of the MIMO channel.

30. The apparatus of embodiment 29, wherein the means for calculating the correlation matrix includes: means for determining the energy of each of the plurality of channel impulse response matrices, and means for calculating a correlation matrix of a channel impulse response matrix with the highest energy among the plurality of channel impulse response matrices, to obtain the correlation matrix for the MIMO channel.

31. A processor-readable means for storing instructions operable to: receive a plurality of channel response matrices for a channel response from a multiple input, multiple output (MIMO) channel in a MIMO system; calculating a correlation matrix for the MIMO channel, based on the plurality of channel response matrices; and decomposing the correlation matrix to obtain at least one guide vector for at least one spatial channel of the MIMO channel, wherein said at least one guide vector is used by a transmitting entity for frequency independent spatial processing of a data flow sent by said at least one spatial channel associated with said at least one guide vector.

32. The processor-readable medium of embodiment 31, and further storing operable instructions to: compute a correlation matrix of each among the plurality of channel response matrices, to obtain a plurality of correlation matrices for the plurality of matrixes of channel response; and summing the plurality of correlation matrices to obtain the correlation matrix for the MIMO channel.

33. The processor-readable medium of embodiment 31, wherein the plurality of channel response matrices comprise a plurality of channel impulse response matrices for a plurality of time delays of a channel impulse response of the MIMO channel.

3. 4. The processor-readable medium of embodiment 33, and further storing operable instructions to: calculate the energy of each among the plurality of channel impulse response matrices; and calculating a correlation matrix of a channel impulse response matrix with the highest energy among the plurality of channel impulse response matrices, to obtain the correlation matrix for the MIMO channel.

35. A method of performing spatial processing in a multiple input and multiple output (MIMO) communication system, comprising: obtaining a plurality of channel impulse response matrices for a MIMO channel in the MIMO system, wherein the plurality of channel impulse response matrices comprise a plurality of time delays of a channel impulse response of the MIMO channel; calculating the energy of each among the plurality of channel impulse response matrices; identifying a channel impulse response matrix with the highest energy among the plurality of channel impulse response matrices, as a channel impulse response matrix for a main MIMO channel path; calculating a channel impulse response matrix correlation matrix for the main path; and decomposing the correlation matrix to obtain a guide vector for a spatial channel of the main path, wherein the guide vector is used by a transmitting entity for frequency-independent spatial processing of a data stream sent via the channel. MIME.

36. The method of embodiment 35, in which the self-value decomposition of the correlation matrix for the main path is performed to obtain the guide vector for the spatial channel of the main path.

37. The method of embodiment 35, further comprising: obtaining a correlated filter for each among a plurality of receiving antennas at a receiving entity, based on the guide vector and a plurality of channel impulse response vectors for the antenna receiving, wherein the plurality of channel impulse response vectors for each receiving antenna are obtained from the plurality of channel impulse response matrices; and filtering a plurality of received symbol streams for the plurality of receive antennas with the plurality of correlated filters.

38. A method of performing spatial processing in a wireless communication system with a plurality of transmitting antennas at a transmitting entity and a plurality of receiving antennas at a

ES 2 525 141 T3 receiving entity, the procedure comprising: obtaining a plurality of sets of channel response vectors for the plurality of receiving antennas, one set for each receiving antenna, wherein each set of channel response vectors is indicative of a channel response between the plurality of transmit antennas and one of the plurality of receive antennas; calculating a correlation matrix for each among the plurality of receiving antennas, based on the set of channel response vectors for the receiving antenna; and decomposing the correlation matrix for each receiving antenna, to obtain a guide vector for the receiving antenna, wherein a plurality of guide vectors for the plurality of receiving antennas are obtained, and the plurality of guide vectors are used by the transmitting entity for frequency independent spatial processing of at least one data stream sent to the receiving entity.

39. The method of embodiment 38, in which the calculation of the correlation matrix for each receiving antenna includes: calculating a correlation matrix of each among the plurality of channel response vectors for the receiving antenna, to obtain a plurality of correlation matrices for the plurality of channel response vectors for the receiving antenna, and summing the plurality of correlation matrices for the plurality of channel response vectors for the receiving antenna, to obtain the correlation matrix for the receiving antenna.

40. The method of embodiment 38, further comprising: obtaining a correlated filter for each among the plurality of receiving antennas, based on the guide vector and the set of channel response vectors for the receiving antenna; filtering a stream of received symbols for each among the plurality of receiving antennas with the correlated filter for the receiving antenna, to obtain a stream of filtered symbols for the receiving antenna; and combining a plurality of filtered symbol streams for the plurality of receiving antennas, to obtain at least one detected symbol stream for said at least one data stream sent by the transmitting entity.

41. The method of embodiment 38, in which a data stream is sent by the transmitting entity to the plurality of receiving antennas, using the plurality of guide vectors.

42. The method of embodiment 38, in which a plurality of data streams are sent by the transmitting entity to the plurality of receiving antennas, using the plurality of guide vectors.

43. The method of embodiment 42, further comprising: obtaining a correlated filter for each among the plurality of receiving antennas, based on the guide vector and the plurality of channel response vectors for the receiving antenna, wherein a plurality of correlated filters are obtained for the plurality of reception antennas; filtering a plurality of received symbol streams for the plurality of receiving antennas with the plurality of correlated filters, to obtain a plurality of filtered symbol streams; and combining the plurality of filtered symbol streams to obtain a plurality of detected symbol streams for the plurality of data streams sent by the transmitting entity.

44. The method of embodiment 34, further comprising: performing space and time equalization on the plurality of detected symbol streams, to obtain a plurality of recovered symbol streams for the plurality of data streams.

Four. Five. An apparatus in a wireless communication system with a plurality of transmitting antennas at a transmitting entity and a plurality of receiving antennas at a receiving entity, the apparatus comprising: a channel estimator for obtaining a plurality of sets of channel response vectors for the plurality of receiving antennas, one set for each receiving antenna, wherein each set of channel response vectors is indicative of a channel response between the plurality of transmitting antennas and one of the plurality of receiving antennas; and a controller for calculating a correlation matrix for each among the plurality of receiving antennas, based on the set of channel response vectors for the receiving antenna, and for decomposing the single correlation matrix for each receiving antenna, to obtain a guide vector for the receiving antenna, wherein a plurality of guide vectors are obtained for the plurality of receiving antennas and the plurality of guide vectors are used by the transmitting entity for frequency-independent spatial processing of at least one data stream sent to the receiving entity .

46. The apparatus of embodiment 45, in which the controller calculates a correlation matrix of each among the plurality of channel response vectors for each receiving antenna, to obtain a plurality of correlation matrices for the plurality of response vectors for the receiving antenna, and for summing the plurality of correlation matrices for the plurality of channel response vectors for the receiving antenna, to obtain the correlation matrix for the respective receiving antenna.

47. The apparatus of embodiment 45, in which the controller obtains a correlated filter for each of the plurality of receiving antennas, based on the guide vector and the set of channel response vectors for the respective receiving antenna.

ES 2 525 141 T3

48. The apparatus of embodiment 47, further comprising: a plurality of correlated filters for the plurality of receiving antennas, one correlated filter for each receiving antenna, and each correlated filter is used to filter a stream of received symbols for the receiving antenna. associated reception, to obtain a filtered symbol stream; and a combiner for combining a plurality of filtered symbol streams from the plurality of correlated filters, to obtain at least one detected symbol stream for said at least one data stream sent by the transmitting entity.

49. An apparatus in a wireless communication system, comprising: means for obtaining a plurality of sets of channel response vectors for a plurality of receiving antennas, one set for each receiving antenna, wherein each set of response vectors of channel is indicative of a channel response between a plurality of transmit antennas and one of the plurality of receive antennas; means for calculating a correlation matrix for each among the plurality of receiving antennas, based on the set of channel response vectors for the respective receiving antenna; and means for decomposing the single correlation matrix for each receiving antenna, to obtain a guide vector for the respective receiving antenna; wherein a plurality of guide vectors are obtained for the plurality of receiving antennas, and are used by a transmitting entity for frequency-independent spatial processing of at least one data stream sent to a receiving entity.

fifty. The apparatus of embodiment 49, further comprising: means for calculating a correlation matrix of each among the plurality of channel response vectors for each receiving antenna, to obtain a plurality of correlation matrices for the plurality of channel response vectors for the receiving antenna, and means to sum the plurality of correlation matrices for the plurality of channel response vectors for each receiving antenna, to obtain the correlation matrix for the respective receiving antenna.

51. The apparatus of embodiment 49, further comprising: means for obtaining a correlated filter for each of the plurality of receiving antennas, based on the guide vector and the set of channel response vectors for the respective receiving antenna; means for filtering a stream of received symbols, for each of the plurality of receiving antennas, with the filter correlated for the receiving antenna, to obtain a stream of filtered symbols for the respective receiving antenna; and means for combining a plurality of filtered symbol streams for the plurality of receiving antennas, to obtain at least one detected symbol stream for said at least one data stream sent by the transmitting entity.

52. A computer-readable medium for storing instructions operable to: receive a plurality of sets of channel response vectors for a plurality of receiving antennas, one set for each receiving antenna, wherein each set of channel response vectors is indicative of a channel response between a plurality of transmitting antennas and one of the plurality of receiving antennas; calculating a correlation matrix for each among the plurality of receiving antennas, based on the set of channel response vectors for the respective receiving antenna; and decomposing the correlation matrix for each receiving antenna, to obtain a guide vector for the respective receiving antenna, wherein a plurality of guide vectors for the plurality of receiving antennas are obtained, and used by a transmitting entity. for frequency independent spatial processing of at least one data stream sent to a receiving entity.

53. The processor-readable medium of embodiment 52, and further storing instructions operable to: compute a correlation matrix of each among the plurality of channel response vectors for each receiving antenna, to obtain a plurality of correlation matrices for the plurality of channel response vectors for the respective receiving antenna; and summing the plurality of correlation matrices for the plurality of channel response vectors for each receiving antenna, to obtain the correlation matrix for the respective receiving antenna.

54. The processor-readable medium of embodiment 52, and further storing instructions operable to: obtain a correlated filter for each among the plurality of receiving antennas, based on the guide vector and the set of channel response vectors for the respective receiving antenna; filtering a stream of received symbols for each among the plurality of receiving antennas with the filter correlated for the receiving antenna, to obtain a stream of filtered symbols for the respective receiving antenna; and combining a plurality of filtered symbol streams for the plurality of receiving antennas, to obtain at least one detected symbol stream for said at least one data stream sent by the transmitting entity.

55. A procedure for performing spatial processing in a multiple input and single output (MISO) system, using orthogonal frequency division multiplexing (OFDM), the procedure comprising: obtaining a set of channel response vectors indicative of a channel response between a plurality of transmitting antennas at a transmitting entity and a receiving antenna at a receiving entity in the MISO system; calculating a correlation matrix based on the set of channel response vectors; and decomposing the correlation matrix to obtain a guide vector used by the transmitting entity for frequency-independent spatial processing of a data stream sent to the receiving entity.

ES 2 525 141 T3

56. The method of embodiment 55, in which frequency independent spatial processing is performed by the transmitting entity in the time domain, on a stream of time domain segments generated for the data stream by OFDM modulation.

57. The method of embodiment 55, in which frequency-independent spatial processing is performed by the transmitting entity in the frequency domain, for each of a plurality of subbands, on data symbols generated for the stream of data.

58. The method of embodiment 55, further comprising: obtaining a correlated filter based on the guide vector and the set of channel response vectors; and filtering a stream of received symbols with a correlated filter, to obtain a stream of detected symbols.