CA2495630C - Spatio-temporal processing for communication - Google Patents

Abstract

The invention provides a space-time signal processing system with advantageously reduced complexity. The system may take advantage of multiple transmitter antenna elements and/or multiple receiver antenna elements, or multiple polarizations of a single transmitter antenna element and/or single receiver antenna element, and is not restricted to wireless contexts. One embodiment employs a substantially orthogonalizing procedure (SOP) in conjunction with a plurality of transmitter antenna elements and one receiver antenna element. The SOP decomposes the time domain space-time communication channel that may have inter symbol interference (ISI) into a set of parallel, space-frequency, SOP bins wherein the ISI is substantially reduced and the signal received at a receiver in one bin of the SOP is substantially independent of the signal received in any other bin. The decomposition of the ISI-rich space-time channel into substantially independent SOP bins makes it computationally efficient to implement various advantageous spatial processing techniques.

Description

wo ~ero~ss rcrrtrsrrnsi6o SPATIO-TEMPORAL PROCESSING FOR
s COMMUNICATION
to 13 The present invention relates to digital communication and more particularly to a space-time co~cation systmcs.
The abgity to conmewnicate tough wireless media is made difficult by the inherent characteristics of how transmitted signals propagate through the emrironazent. A
commmncation signal transmitted through a ttaasmitter antenna element travels along 20 multiple paths to the receiving antenna element. Depending on many factors including the ~gnal frequency and the terrain, the paths along which the signal travels will exlubit different attenuation and propagation delays. This resorts in a cornanuiication channel vvbich exhibits B~~Y
It is well known ttrat adaptive spatial Processing using multiple as~nna arrays 25 increases the communications quality of wireless systems. Adaptive array Processing is known to improve bit error rate, data rate, or spectral eff scary in a wireless communication system. The prior art provides for methods imrolving some form of space-time signal proc~aing at either the input to the channel, the output to the cbatmsl, or both. ?he space-time Processing step is typically acxoarplished use an equalization strucaire wlea~eia the 30 time domain equalizer tap settings for a multitude of antennas are simultaneously optimized.
This so-called "space-time equalizaion" leads to high signal processing complexity if the delay spread of the equivalent digital channel is substantial.
There is prior art teaching the use of com~ional antenna beams or polatizations to create two or more spatially isolated communication charmels betvvxa a 35 transmitter and a receivar, lrut only under certain favorable conditions.
The radiation pattern _ cross talk between different physical transmit and receive antenna pairs must provide su~aent spatial isolation to create two or more substant;ally independent communication channels. This can lead to stringent manufacturing sad performance requirements on the _ physical artterma strays as well as the receiver and tran~nitter electronics. In addition, when 40 large objects in the wireless propagation channel cause muttipath reBections, the spatial isolation provided by the prior art between say two spatial subchannels can be severely - degraded, thus reducing communication quality.

wo ssro~3ss rc~rrusmnst6o

2 What is needed is a system for more effectively taking advantage of multiple transmitter antennas andlor multiple receive antennas to ameliorate the deleterious eats of the inherent characteristics of wireless media.
SUMMARY OF THE INV~'.NTION
The present imrention provides a space-time signal processing system with advantageously reduced complexity. The system may take advantage of multiple transmitter antenna elements and/or multiple receiver antenna elements, or multiple polarlzations of a single transmitter antenna element and/or single receiver antenna element. The system is not restricted to wireless contexts and may acpioit any channel having multiple inputs or multiple outputs and certain other characteristics. In certa'sn embodiments, mufti-path effects in a transmission medium cause a multiplicative increase in capac'tty.
One wireless embodiment operates with an efficient combination of a substantially orthogonalizing procedure (SOP) in conjunction with a phualit3r of transmitter arnenna elements with one receiver antenna element, or a plurality of receiver antenna elements with one transmit antenna element, or a plurahxy of both transmitter and receiver antenna elements. The SOP decomposes the time domain space-time conununication chancel that may have inter symbol imerference (ISl) into a set of parallel, space-frequency, SOP bins wherein the ISI is substantially reduced and the at a recdva in one bin ofthe SOP is substantially independent of the signal received in any other bin of the SOP. A major benefit achieved thereby is that the decomposition of the ISI»rich space time channel into substantially independent SOP bins makes it computationally efficient to implement various advantageous spatial proces~g techniques embodied herein. The e>ficiency benefit is due to the fact that the total signal processing complexity required to optimize performance is all of the SOP bins is often significantly lowEr than the processing compladty required to jointly optimize multiple time domain equalizes.
Another benefit is that in many types of wireless channels where the rank of the matrix channel that exist bewe~ the transmitter and the receiver within each SOP bin is greater than one, the combination of an SOP with spatial processing can be used to e~ciently provide multiple data conunuriication subchannds within each SOP bin. Tlns has the desirable effect of essentially multiplying the spectral data effiaency of the wireless system. A
further feature is the use of spatial processing techniques within each transmitter SOP bin to reduce radiated irn~ to uninttritional receivers. A still further featiue is the ability to perform spatial processing within each receiver SOP bin to reduce the deleterious effects of itrtafereace from uniatmtional transmitters.
One advantageous specific embodiment for the SOP is to transmit with IFFT
basis functions and receive with FF'f basis functions. This particular SOP is commonly referred to as discrete orthogonal frequency division multiplexing (OFDM), and each SOP bin is thus associated with a frequency bin. This tmboduneM enhances OFDM with the addition of e~cdent spatial processhig technicp~.
According to the present imrention, space-frequency processing may adaptively create substa~iauy independent spatial subclumnels within each SOP
bin even in the presence of significant cross talk imerfereace between two or more physical transmit and r~eive anterma pairs. A further advantage is that the space-frequency procxssing can advantageously adapt to cross talk interference between the phytical antenna pairs even if this cross-talk is frequency dependent, or time varying, or both. Thus, the present invention may provide two or more substantially independent communication channels even in the preaen<x of severe inunipath and relatively poor physical antenna radiation pattern performance.

wo 3ss ~ pcrnrs~natso

3 A further uaderstaading of the nature and advamagea of the imrentions herein may be realised by reference to the remaining portions of the specification and the attached drawings.

wo ~3ss rc~rrosr~rnsiso BRIEF DESCRIPTION OF TSE DRAWINGS
Fig. 1 depicts a transmitter system according to one embodiment of the present invention Fig. 2 depicts a particular substantial orthogonaiizing procedure (SOP) useful in one embodiment of the present invention.
Fig. 3 depicts a receiver system according to one embodiment of the present ikon.
Fig. 4 depicts a first comrmmication scenario where multipath is found.
Fig. 5 depicts a second ~nummication scenario where multipath is found.
Fig. 6 depicts a third communication sconario where multipath is found.
Fig. 7 depicts a multiple-input, muhipIe-output (MIMO) channel with intaf~cee.
Fig. 8 depicts the use of an SOP in a single-input single-output (SISO) channel.
F'ig. 9 depicts the use of an SOP in a MQI~iO channel according to one embodiment of the present invention Fig. 10 depicts the operation of an SOP in the context of one e~nbodimem of the present ion Fig. 11 depicts the application of spatial processing to a particular SOP bin at the transmitter end according to one embodiment of the present invention.
Fig. I2 depicts the application of spatial processing to a particular SOP bin at the receiver end according to one embodiment of the present invention.
Fig. 13 depicts the application of spatial processing to N SOP bins at the transmitter end according to one embodiment of the present invention.
Fig. 1~ depicts the application of spatial processing to N SOP bins at the receiver end according to one embodime~ of the present invention.
Fig. 15 depicts the use of a single spatial direction at the transmitter end for each bin of an SOF according to one embodiment of the present imrention.
Fig. 16 depicts the use of a single spatial direction at the receiver end for each bin of an SOP according to one embodiment of the present imrentioa Fig. 17 depicts the use of one or more common spatial weighting vectors far all SOP bins at the transmitter end according to one embodiment of the present imrentioa Fig. 18 depicts the use of one or more common spatial weighting vectors for all SOP bins at the receiver end according to one embodiment of the present invention.
Fig. 19 depicts the use of an encoder for tech SOP bin according to one embodiment of the present invention.
Fig. 20 depicts the use of an encoder for each spatial direction according to one embodia~nt of the present inve~ion.
Fig. 21 depicts the use of an encoder for each spacef&~equency SLCbchaanel according to one embodiment of the present invention Fig. 22 depicts dishibution of encodes output over all spacelfrequ~y .
subchanaela according to one embodiment of the present invention.
Fig. 23 depicts a detailed diagram of an encoderru~terieaver system according to one embodiment of the present invention.
Fig. 24 depicts a transmitter system wherein multiple spacelfrequency subcharmels are employed without spatial orthogonalization according to ono embodiment of the present invention.

wo 3ss rcr~crs~nsiso Fig. 25 depicts a raxiver system wherein multiple spacelfrequency subchannels are employed without spatial orthogonstlizatton according to one embodiment of the presort invention.
Fig. 26 delricts an exemplary technique for bit loading with a trellis coder that 5 uses a one-dimensional QAM symbol constellation.
DESCRIPTION OF SPECIFIC EMBODIMENTS
Definitions A "channd" refers to the inert symbol to output symbol relationship for a communication system. A '5rxtor channel" refem to a charind with a single input and multiple outputs (SIMO), or multiple inputs and a single output (MISO). Each h; entry in the vector channel h describes one of the complex path gains present is the chaand. A "matmc channel" refers to a e~anad with multiple inputs and multiple ~tputs (N1n140).
Each entry H;,, in the matrix H describes the complex path gain fi-om input j to output i. A "space tune channel" refers to the input to output relationship of a ~O matrix channel, or a SIMO or MISO vector channd, that occurs when multipath sigml propagation is present so that the channd contains dday ddnents that produce inter-symbol intafet~enca (ISn as explained blow.
A "spatial direction" is a one dimensional subspace within a matriJC or vector communication chanad. Spatial direcxions need not be arthogonet. A spatial direction is typically chsractuized by a complex input vector and a complex output vector used to traasmitted or revived signals as eo~plained heron.
A "sub-channel" is a combination of a bin in a substantially orthogonalizing procedure 23 (SOP) as explained blow and a spatial direction within that bin. A group of spatid subchanaels within an SOP bin may or may not be orthogonal.
An "orthogonal dimension" is one member in a set of substantially orthogonal spatid directions.
A channd "subspace" is a characterization of the wmplex m-apace direction occupied by one or more m-tonal vectors. The subspace chazacterization can be based on the instantaneous or average behavior of the vectors. A subspace is oRen character by a vector-subspace of a covariance matrix. The covariance matrix is typically a time or frequency averaged outer product of a matrix or vector quantit5r. The covariance matrix characterizes a collection of avaage channel directions and the assoaated average strength for each direction.
A "two norm" metric for a vector is the sum of the squared absolute values for the dements of the vaxor.
A "Euclidean metric" is a two norm metric.
_ "Intersymbol interference" (ISl) refers to the self interference that occauu~s between the delayed and scaled versions of one tiny domain symbol and subsequent symbols received at the output of a delay spread com<rumicaiion cl>annd. The chaand delay spread is ceased by the digerence in propagation delay between the various muItipath components combined with the time domain response of the RF and digital filter el~neats.
A "substantially orthogonalizing pmaedure" (SOP) is a procedure that plays a part in - 45 transforming a time domain sequence ~to a paralld set of ~bstantially orthogonal bins, wherein the signals in one bin do not substantially interfere with the signals from other bins.
Typically, the transformation from a time domain sequence to a set of substantially orthogonal bins requires a tran~itter SOP with a set of input bins, and a r SOP with a set of output bias.
"Convolutional bit mapped QAM" (CBM-QAM) is the coding system that results when the output of a convoiutional encoder are grouped and mapped to QAM
constellation points.
Fading "stiuc~ure" occurs why t6e fading behavior of one or more entries in a channel matrix within an SOP bin is cowelated across time, or frequency, or both. This strucbrre can be exploited using advantageously designed estimation filters to improve channel estinnation accuracy given multiple frequency samples of the channel matrix entries, or multiple time samples, or both.
A "tnaximum likelihood sequence detector" is a sequmcx Bator that computes the most likely transmitted code sequence, from a set of possible sequences, by minimizing a mmdmum likelihood cost firaction.
An "a~enna element" is a physical radiator used to transmit or receive radio frequency signals. An antenna element does not involve any electronics processing components. A single radiator with two polarization feeds is viewed as two antenna elements.
An "alma array" is a collection of antenna elements.
A "burst" is a group of transmitted or received communication symbols.
Background Mat~isl The disclosure herein assumes a background in digttal communication and linear algebra. The following ~ are incorporated lain by rofenmce.
Wozencraft & Ja,cobs, Principles of Communication Engineming (1965).
Fiaykin, Adaptive Filter Thcory, 2"~ Ed. (1991).
Strung, l~ar Algebra, 3'~ Ed. ( I 988).
fakes, Ivhicrowave Mobile Communication (I974).
Proakis, Digital Communications (1995).
Transmitter Ovavisrw Fig. 1 depicts a transaritter system in accordance with one embodiment of the preset invention. Typically an information signal input 2 includes a digital bit sequ~ce, although othex forms of digital data or analog data are poss~'ble. In the cage of digital data, the input data sequence is f rst fed into an Encoder and Interleaving appin~atus 10 where the data is encoded into a symbol stream. The symbol stream is typically a sequence of complex 3 S digitized values that represent members of a furite set. Each symbol can be a one dimensional value, or a multidimensional value. An exemplary one dimensional symbol set is a PAM
constellation. Note that is this discussion, it is understood that a symbol with in-phase and quadrature camponems, is considered to be a complex one dimensional symbol; so that the QAM constellation is also viewed as a set of one dimensional symbols. An example multidimensional symbol set is a sequential grouping of QAM constellation members.
The purpose of the encoding process is to improve the hit error rate of the transmhted signal by irn~cing some form of information redundancy imo the transmitted data strsam.
Useful encoding techniques can involve combinations of a number of well known techniques such as comrolutional encoding with bit mapping to symbols, trellis encoding, block coding such as cyclic redundancy check or Reed Solomon coding with bit mapping or Automatic Repeat Queing. An interleaves is often advantageous for distributing the transmitted i~ormation among the various subchannels available for transmission. This imerleaving distributes the effects of channel fading and interference so that long sequences of symbols wo 9sro9~ rcrms~~mn6o with poor quality are not grouped closely together in the SOP bin s~u~Ce that is fed into the receiver decoder. In many applications, it is advantageous to perform a power and bit-loading opt>mization the arunber of bits that are mapped to a given encodex symbol, and the signal power assigned to that symbol, are determined based upon the measured communication quality of the space-frequency information subcharmel that carries the symbol Afler the digital data is encoded into a sequence of symbols, a Training Symbol Injection block 20 may be used to place a set of known training symbol values in the transmitter symbol stream. The purpose of the training symbols is to provide a known input within a portion of the transmitted symbol stream so that a receiver may estimate the communication channel parameters. The channel estimate is used to aid in demodulation and decoding of the data sequence. The training symbols may be injected periodically in time, periodically in frequency, or both. It will be obvious to one skilled in the art that blind adaptive spatial processing techniques can be utilized within each SOP bin at the receiver as an alternative to training with known symbols. In such blind detection implementations, Training Symbol Injection block 20 is unnecessary.
The data plus training symbol steam is then fed into a Transmitter Space-Frequency Pre-Processor (TSFP) block 30. The TSFP block 30 performs two sets of advantageous processing steps on the symbol stream before transmission. One procxssing step 2U accomplished within the TSFP is the transmitter portion of a substantrall3r orthogonaIiziag procedure (SOP). When the transmitter portion of the SOP is combined with the receiver portion of the SOP, a set of parallel bins are created in such a manner that information transmitted within orar bin does not substantially interfero with information transmitted from another bin after the receive portion of the SOP is completed. One pSOP pair is the inverse fast Fourier traoaform (IFFT) at the transmitter combined with the FFT
at the tooeiver. Another advantageous SOP pair embodiment is a bank of n~riple filter and frequency converter pairs (multi~and SOP) with one filter bank located at the transmitter and one filter hank located at the receiver as depicted in Fig. 2. Several other example SOPS
including the ~ylbert transform pair and generalized wavelet transform pairs will be obvious to one skilled in the art. The other processing step accomplished in the TSFP
is spatial Prooess~g- The spatial processing step typicalhr ~Phes one or nmre symbols that are dfor transmission in a given SOP bin with one or more spatial vector weights.
For conveni~ce in the following discussion, the collection of spatial processing weights applied to the signals transmitted or received in a given SOP bin are sometimes referred to es a 3 S matrix. The spataal vector weights are optimized to obtain various desirable paformaace enhancements.
Fig. 2 depicts a digrtal baseband fr>ter bank at the reoavec and a filter beak located at the transmitter. Each filter of the transmitter filter bank includes a mixer (frequency converter) 60, a bandpass iiher 70 and as interpolator 80. Each filter of the receiver fitter bank includes a barrdpass fiher 90, a mixer 60, and a decimator 100.
One transmitter embodimern optimizes the transmitter spatial vector weights so that the multiple subchannels in a given SOP bin can be converted at the receiver into subatantislly indept raved spatial subchaands wherein symbols from one subchanmel do not substantially interfere with symbols from another subchaanel. Another embodiment optimizes the transmitter spatial vector weight to improve the received power of one or more spatial subchannels within each SDP bin, or to innprove the average power of savaal spatial subchannels within several SOP bins. A further embodiment optimizes the transmitter spatial _ vector weights within each SOP bin to simultaneously incraise the powe~c delivered to the desired receiver within one or more spatial subchannels while reducing interference radiated wo ~aro93ss ~ Pcr~us~rnsiso to unintended receivers. A yet further emb«iimear spatial processes one or more symbols within each SOP bin by multiplying each symbol with a transmitter weight vector that is fixed for all SOP bins, with the weight vectors optim;zed to increase the time or frequency average power delivered to one or more desired receiver spatial subchands, and possz'bly reduce the time or frequency averaged interference radiated to unintended receivers. This last embodiment is particularly useful in FDD systems where multipath fading makes it impossible to estimate the forward channel from reverse channel data, but where the average forward chamrd subspaces are substantially similar to the average reverse channel subspaces. Another embodiment teaches simply routing each symbol from the encoder to one antenna element in each SOP bin without any weighing. Other useful embodirn~ts are discussed herein, and many others useful combinations of spatial processing with an SOP will become obvious to one skilled in the err. It is undcrstaod that one or more digital filters are typically used izr TSFP 30 to shape the transmitted RF signal spectnim.
Once the encoder symbol sequence is processed by TSFP 30, the processed symbol sequence includes a parallel set of digital time domain signal sequences. Each of these time domain signal sequences is fed into one input of a Modulation and RF System block 40.
Modulation and ItF System block 40 includes a set of independent RF upcomrater chains that frequency convert the digital baseband signal stquence up to the RF carrier frequency. This is accomplished using apparatus that idigital to analog comrdters, RF mixes apparatus, and frequency synthesizer apparatus. The details of these elements of the invention an well known and will tmt be discussed here.
The final step in the transmission process is to radiate the transmitted signal using a Transmit Arrteona Array 50. The saterma arrays can be consducted from one or more oo-polarized radiating elements or there may be multiple polarizations. If there is multipath 23 signal propagation present in the radio link, or if there are multiple potarizations in the ar>tenna arrays, or if at least one of the antenna eie~ments on one side of the link are in a disparate location from the other elements on the same side of the link, then the imr~ion has the advantageous ability to create more than one subchannel within each SOP
bin. It is understood that one physical antenna reflector with a feed that has two polarizations is considered as two mitenna elks in all that follows. There are no restrictions on the antenna array geometry or the geometry of each radiating element. A
t<anstnitter system invention may adapt to provide optimised performance for arty arbitrary antenna array.
Radvex Overview Fig. 3 depicts a receiver system according to one embodiment of the presern invention. The RF signals from each of the dements of an Antenna Array 110 are downcwrvated to digital basebaad u~ag a Demodulation and RF System 120.
Demodulation and RF System 120 includes the RF signal processing apparatus to downcorrvert the 1tF
carrier signal to a baseband IF where it is then digitized. After the digitizer, a timing and frequency synchronization apparatus is used to recover the timing of the tran~nitted digital signal sequence. Several known techniques may be used for the purpose of synchronization and these techniques will not be discusa~ed heteicr.
In certain embodiments of the invention, a8er Demodulation and RF System 120, the digital bssebaad signal is then fed i~o a Channel ID block 130 and a Receiver Space-45, Frequency Processor (RSFP) block 140. Within Channel ID block 130, the ci~aracteristics of the digital communication channel are estimated. The estimated channel values consist of entries in a martyr for each SOP bin. The matrix contains complex values rGpt~tting the magnitude of the spatial channel within the SOP bin from one transmit antenna eleme~ to one wo 3ss rcrmss~nsi6o receive antau~a danent. The mavix chanad estimate for tech SOP bin is providai to RSFP
block 140 and Decodes and Deimtrleaving block 150.
Some embodiments of the invention involve improving chaand estimation performance by exploituig the structured nature of the frequency domain fading that exists in the matrix channels across SOP bins, exploiting the structure in time domain fading of the matrix channds, or exploiting both the frequency and time domain fading structure that is present. By exploiting the frequency domain fading correlation, the entire set of matrix channds within the SOP bins may be estimated even when trw>;ag information is transmitted in a subset of the SOP bins. This allows for simultaneous transmission of training and data thus reduang overhead. By axploi'g the time domain correlauan of the c6annd fading within each SOP bin, channel e~nation accuracy is increased for a given time epoch between training events. This reduces the required frequency of training symbol transmission and thus further reduces training overhead. It is understood that it is also poss~'ble to separately exploit fink domain and frequency domain correlation, with the most beneficial results occurring if both correlation dimensions are used advantageously. It is to be understood that Charnel 1D
block 130 is shown as a separate function evens though it may share some dements with ItSFP block 140 or Decoder and Deinterleaver block 150.
RSFP block 140 performs the receiver signal processing that is the dual of the two sets of operations performed in TSFP 30. One of the steps performed in RSFP
140 is the ra~iver half of the SOP. As discussed about, the receiver half of the SOP
completes the transformation between the time domain channel with ISI to tin aubst~tidllr orthogonal set of bins. The second set of signal processing operations that can be performed in the ItSFP is spatial proces~ng. In one class of embodiments, the receiver spatid processing step combines the output of the SOP bins using one or more vector weighted inner product steps to form one or more one-dmaa~sional received spatial subchannels within each SOP bin. The receiver weight vectors are chosen to optimize an advantageous performance measwe. In one embodiment, whmnia both the transmitter and receiver have knowledge of the chamm!
state information within each SOP bin, the traoarrntter spatial weight vectors and the receiver spatial weight vectors are bath chosen to optimize performance for a set of rrmltiple substantially independent subchannels within each SOP bin. As discussed above, this sigrd5cantly increases the spectral ei~ciency of the system. In another eZnbodiment wherein the transmitter does not have channel state information, the receiver performs the spatial proce~ng required to create raultiple substantially independent subchannda within each SOP bin. In a further embodiment wherein the otter may or may not have channel states information, the receiver reduces the effects of interference radiated from udatentioml otters as well as p~foraung the spatial processing required to seats multiple substantially independem subchannels within each SOP bin. A yet further embodiment optimi~ the receiver spatial vector weights within each SOP bin to simultaneously increase the received power and reduce the detrirnemal effects of interference roceive~d from urdateutiond transmitters. An additional embodiment irrvolves forming one or more vector weights, that are fitted for all SDP bins, where the vector waghts are optimized to simultaneously increase the time or frequency averaged received power for one or more spatial subchannels, while possibly also reducing the time or frequency averaged inte~~n~
power received from unintended transmitters.
As discussed herein, certain embodiments involve simply passing the vector samples received in each SOP bin to Decoder and Deinterleaving block 150 without performing any spatial processing. It will be obvious to one stalled in the art that other combinations of transmitter spatial weight vector optimizatian techniques and receiver spatial weight vector optimization techniques can be constructed around the principle concept of spatial processing wo rcr~nsi6o IO
in combination with an SOP. Other embodiments are discussed herein. One experienced in the art will be able to recognize additional embodiments that involve advantageous combination of an SOP with spatial processing at the raxiver or the transmitter. It is understood that one or more digitd filters are typically used in RSFP block 1400 to shape the received RF signal spectrum.
The outputs of RSFP block 140 are fed into Decoder and Deiaterleaving (DD) block 150. There are two broad exemplary classes of operation for the DD block 150.
In the first exemplary broad class of embodiments, DD block I50 decodes a symbol sequence which was encoded and transmitted through a multitude of SOP bins with one or more substantially independent suds. The decoder includes the appropriate receiver counterparts for the combination of encoders selected for the trans:mtter. A prefen~ed embodiment includes a deir~terleaver, a trellis decoder or comrolutional bit mapping decoder employing a scalar weighted Euclidean maximum likelihood sequence detector, followed by a Reed Solomon decoder, followed by an ARQ system to correct Reed Solomon codeword errors. In the second exemplary broad class of embodiments, DD black 150 decodes a sequence of multidimensional symbols, or groups of adjacent one dimensional symbols, with each multidimensional symbol or group of one dimensional symbols being received in an SOP bin.
Typically, the symbol sequencxs are transmitted without weighting or with weighting that optimizes some measure of average signal quality.
In an alternative embodimem, trellis encoded symbols are grouped and interleaved in a manner such that the symbols transmitted from the ant~na dues within a given SOP bin form a vector that is drawn from either a multidimensional QAM encoder output symbol, or a sequence of one dimensional QAM e~oder outfit symbols that have adjacart locations in the pre-interleaved encoder output sequence. In this way, a maximum likelihood vector decoder may be constructed given an estimate of the channel matrix that is print within each SOP bin. The maximum likelihood decoder computes the weighted vector Euclidean metric given the deintaleaved r~dv~ vector firm each SOP bin, the deit~terleaved maxmc channel estimates from each SOP bin, and the transnvtted vector symbol trellis state table.
In another alternative embodiment, either of the aforememioned encoder embodiments will have preferable performance if the encodes polynomial and symbol constellation set are optimized to improve the bit error rate performance given the characteristics of the matrix channel fading that occurs in each SOP bin. One particmlar metric that is well suited for a code polynomial optimization search is the product of the two norms of the vector difference bexween the correct transmitter symbol vector and the error symbol vector.
The output of DD block 150 is the estimatad bit stream at the raxive end of the radio link.
It is to be understood that all transmitter embodiments of the pn~ent invention may be adapted for use with a receiver accessing the channel through a single ehanud output such as a single receiver antenna element. Furthermore, all receiver embodiments of the present imrention may be adapted for use with a transmitter accessing the channel through a single channel input such as a single transmitter antenna element. It is understood that the channel is then a vector channel. Such multiple-input single-~tput (NBSO) and single-input zxnrltiple output (SIMO) systems are within the scope of the present invention.
Space-Frequency Communication in a Multipath Chanad Before dcwaloping the signal processing of the present invention, a physical description and mathematical description ofwirdess channels are provided. Many wireless wo 3ss rc~r~nnsiao communication clwraels are characterized by multi-path, where each path has assoeiatod . fading and propagation delay. Multipath may be .creatod by reflections in the physical radio path, multiple antenna polarizations, antenna elements located in disparate locations, or a combination of arty of these. One scenario in which multipath is created is illustrated in Fig 4.
A Base 152 its information to and receives information from a remote unit I?OA
or I?OB. Base 152 possesses ors3 or more antenna elements referred to as an easy 55,.
Similarly, the Remote Units possess their own sways 55. A transmitted signal propagates along multiple paths 155A-C created by reflection sad scattering from physical objects in the terrain 160A D.
I O Multipath sigma propagation such as that depicted in Figs. 4-6 can give rise to spatially selective fading, delay spread, frequency fading, and time fading.
Spatial fading occurs as the various wavefronts arriving at the roceiver from different propagation paths combine with constructive and destructive interfemnce at different points in space. An ac~tenna array located within this spatially selective field will sample the Seld at various locations so that the signal strength at each array element is different.
Delay spread occurs due to the differing propagation path lengths. The channel delay spread gives rise to a frequency selective digital communication channel at each antenna element.
This frequency response is different for each elemart by virtue of the frequency dependent spatial fading. Finally, if wither the transmitter, or the receiver, or objects in the terrain are moving, the frequency selective spatial fading m71 vary as a function of time. The present imrention is unique in that it is capable of e~ciently and economically adapting to the time varying apacx-frequency channel response to make advantageous use of the inherent properties of such channels.
For several decades, the primary focus of the prior art has been to somehow mitigate the effects of the multipath chacmd. This conventional approach is ill-advisad since multipath c6atnuls give rise to a multiplicative capacity by virdue of the fact that the multi~th induces a rack greater than one in the matrix channel present in each SOP bin.
This providts opportu~r to form multiple parallel subchannels for communication within each SOP bin.
Thus, one should utilize munipath to improve comnumication performance rather than attempt to mitigate its mss. A substantial advatrtage ~ovided by the present inv~tion is the ability to e~ciently and economically exploit the inhaeat capacx<y advantages of multipath channels using a combination of an SOP and spatial processing or spatial coding.
No other strucdu es are knows to effiaently exploit this fundamental adva~age in the preserxe of substantially frequency selective multipath channels.
Fig. 5 illustrates another wireless cbannet scenario in which muttipath is present. A
Base 152 with an antenna array 55 transmits information to and receives information from a Remote 1?OC with an antenna array 55. In this case, both Base and Remote antenna arrays 55 both have eleme~s with differing polarizations. Thus multipath signal propagation acists even if there are no significant reflections in the physical emrironment. The direct line of sight paths 155B, 155E each corresgonding to one of the polarizations in the array elements, are suffcient to create a matrix channel with a rank greater than one within each SOP bin, even if the other reflected radio paths 155A, I55C, 155D, and 155F are insignificant or nonexistent.
It is known that such line of sight polarization (reflection &ee) channels can be decomposed into two parallel communication channels by using high performance dual polarizarion antennas, one at each end of the radio fink, in combination with high performance receiver and transmitter electronics apparatus. In this prior art, the quality of the parallel communication channels is limited by the degree to which the two poiarizafion channels remain independent. In general, maintaining the manufac.~turing tolerances and installation alignment precision in the antennas and electronics required to achieve substantially wo ~3ss rcrnrs9~nsi6o orthogonal spatial subchanneis at the output of the physical receive antenna is relatively expensive. Slight marwfact<uing errors and component variations can lead to s significant cross-talk intaf~se~ between the nwltiple polarizations presa~t in the radio channel. In contrast, an advantageous feature of the inv~tion is that the different polarizations present in the wireless channel may have an arbitrary degree of cross-talk interfereucx, and the cross-tallc intafer~oe may be frequency dependent without loss of performance. In such cases, tlm invention provides an oconomical and edBcient method for fully exploiting the multi-dimensional nature of the multiple polarization channel. It is understood that the invention can provide fiurther capacity advantage if the multiple polarization channel also has reflected sigosl paths. This additional muhipath results in an additional in~ase in the channel matrix rank in each SOP bin that can be further exploited to improve the capacity of the channel.
Fig. b depicts another wireless communication scenario in which nrultipath is present and can be exploited to create multiple dimensions for communication. In Fig.
6, two Bases 152A and 1528 with antenna arrays 5$ communicate with Remote Units 170A and 1708 that also possess antenna arrays 55. In this case, the composite channel is defined as the MIMO
channel baween the antennas of the two Bases 152A and 1528 and the antennas of the Remote Units 170A and 1708. Note that this channel includes direct line of sight paths 1 SSB
and 153B as well as the reflected paths 155A., I SSC, 153A, 153C, and 153D. By virtue of the spatial separation b~ween two Bases 152A and 1528, even if the reflected paths are i~ior nonexistent, this charmd contains multipath that can be exploited using the imrenbon. In addition, the channel from the antennas of the two Bases to the antennas of one Remote is again a matrix channel, with rank greater than one, within each SOP
bin so that multiple parallel dimensions for transoo~ission may be created. In these types of applications, the presem invention provides for the ability to reduce interference radiate to unirnerrtional receivers. Furthermore, the present invention provides the capability of reducing the detrimental effects of revived intafw~ce from unintentional transmitters.
Thus it can be seen that multiple transmitter antenna elements or multiple antenna elements may be either co-located or be found at disparate locations.
The following symbol channGt model applies to all of the above multipath nsdio propagation cases ~ by Figs. 4-6. The channel impulse response includes the s of the P~Pon envirorunent, as well as the digital pulse shaping filters used in TSFP
30 and RSFP 140 , the analog filters used in Modulation & RF System 40 and Demodulation & RF
System 120. Due to the difference in propagation delay between the various multipath components combined with the time domain response of the RF and digital filter elements, a 3 5 single symbol transmitted into the channel is received as a collection of delayed copies. Thus, delayed and scaled versions of ocye symbol interfere with other symbols. This self irner~Oe effect is termed intersymbol imerference or ISI. The delay spread parameter , denoted by v, is the duration in symbol periods of the significant portion of the channel impulse response.
As the uam~tted symbol rate is increased or as the phy~cal goometries in the channel become larger, the delay spread can become so large that conventional space..time processing systems become highly complex. An advantage of the present invention is that the signal processing complexity is relatively low even when the delay spread becomes actremely large. This allows for the economical application of MIMO space-frequency proce~ag techniques at high digital data rates. This effiaent use of signal processing comes about because the invention allows the space-time channel to be treated as a set of substantially independent spatial subchannels without sacrificing channel capacity. In contrast, conventional approaches either attempt to equalize the much more complex space-time channel or alternatively sacrifice capacity.

wo ~~ PCTlUS97l13160 The channel is modeled as time-imrariant over the time spawned by a burst of N
data symbols, but varying from one burst to another. This block time imrariant assumption produces a channel model that is sufficiently acauate for channels wherein the block duration is short compared to the channel fading, or (N+2v~?' « A~, where A~ is the correlation irnerval'. Other models are available wherein the channel varies continuously, but thex models add unnecessary complication to the present discussion. It is understood that rapid time variation in the channel can be another motivation for choosing one of the other SOP
alternatives in the presence of fading rates that are rapid with respect to the burst frequency.
One skilled in the art will be aware of the pertinent issues for a gives application. For example, Orthogonal Frequcucy Domain Multiplexing (OFDM) is an SOP composod of an 1FFT and cyclic prefix as the transmitter SOP, and an FFT as the receiver SOP.
With OFDM, one pertinent issue is frequaiicy domain inter-carrier interference (ICI), which can occur in OFDM systems with extremely rapid fading. Such pertinent issues shape the appropriate choice of channel models for various SOP basis fimcrions.
With this background discussion, it can be verified by one skilled in the art that the relationship between the transmitted burst of basebaad symbols and the received burst of baseband samples may be adequately acpress~ as the spacx-time equation, :(k) = G(k~(k)+I(k), where the index k represents bursts. The composite channel output for owe burst of data, =(k) , is written with all time samples appear in sequence for every receive anterma 1 to MR , =(k) = y (1) ~ . . xt (N + v -1) .. . xul (1) . . . x~~ (N + v -1)~T .
Likewise, the input symbol vector is written, Z(k) = lZi (1) ... Zt (~ ~.. Zur ~) ~.. Zyr (~~ Il r The quamityr I(k), defined the same as z(k) and z(k), represe~s both noise and into.
The MIMO channel matrix, G(k), is composed of single-input single-output (SISO) sub-blocks, G~a (k) ... Gtx, (k) ..
G(k) _ . . a C''°'''~""~"'~ . (1) Gar,.t (k) ... G,~"~Xf (k) Furthermore, each of the SISO sub-blocks, G,,I (k) , is a Toepfrtz matroc describing the input-output rehttionshup between the transmitted symbol burst and the received symbol burst for antenna fit (i,~. This MIMO spaco-time chacmel is iih>stratod by Fig. 7, which shows SISO
sub-blocks 180A D and the addition of inte:ferettce for each receiver sample.
Space-Frequency Processing Embodiments of the present imrention uses space-ficquency processing at either the transrrdtter or receiver, or both, to create effective communication systems in wireless channels. Generally, the processing substarrtsally eliminates the ISI caused by the channel correlation across space (antenna corrda~tion) and time (delay spread)- This pmcessing greatly simplifies the design of the remaining functions that comprise a complete ' The comolattori interval here is defined as the time period required for the fading parameter time-autocorretarion function to decrease to some fraction of the zero-si~i&
value.

wo mss rcrios~rnsi6o communication system, including coding and modulation. Furthermore, the processing approach is based upon a capacity-achieving structure for the MIMO wireless communication channel. Space-frequency processing is composed of one or more of the following: an SOP, a transmit spatial processor, and a receive spatial processor.
Snbsts~atiaUy Orthogonalizing Procedure The use of an SOP in a SISO channel is considered first in order to dhrstrate the invention's ability to eliminate ISI across space and time. The SOP is composed of signal processing operations implemented at both the transmit and roceivc sides of the channel. This is illustrated in Fig. 8 where a Transmitter SOP processor 190 and a Receiver SOP processor 200 jointly perform a completo SOP. The SOP ensures that the N input symbols, in bin 1 through bin N, are transoaitted through the channel in such a way that each output symbol is substantially influenced by only the input symbol of the same frequency bin.
For example, the input symbol in bra 1 is the only symbol to have substantial influence on the output symbol 1 S value in bin 1.
This concept gene~izes to the MIMO system as shown in Fig. 9. For the MIMO
system, each transmitter antenna 51 is preceded by one of Mr identical Transmitter 50P
processors. Likewise, each receiver antenna 111 precedes one of Ma identical Receiver SOP
processors. Hence, the processing path for airy transmitter-receiver antenna pair contains a jointly performed complete SOP. In other words, there exist MaMr SISO SOPs in the MIMO system. By acploiting the property of superposition, this collection of SISO SOPS
comprise a MIMO SOP where any two symbols communicated in different bins exhibit substantially crosstalk, irrelevant of the antennas by which the symbols were transmitted and received. Therefore, the SOP establishes N substamialiy independent MIMO
spatial channels.
Marry different SOP implementations exist, including an 1FFT FFT pair, a bank of multiple narrow-band filters, and generalized wavelet transform pairs. One advantageous example of an SOP is the use of a frequency transform combined with a burst cyclic prefnc application procedure processor 207, as shown in Fig. 10. There is also a cyclic prefix . removal procure processor 206 at the receive end. When the freque~y transform is an IFFT-FFT pair 20S and 208 as shown in Fig. 10, this particular SOP is commonly referred to as discrete orthogonal frequency division multiplexing (OFDM). Hence, this embodiment of the invention combines the OFDM SOP with multiple input antennas, or multiple output antennas, or both narltiple input and multiple output antennas. The present embodinnerit is 3S thus termed matrix-OFDM (MOFDM).
The analy~s preserrted for MOFDM will have a substama~ly similar form as other choices for the SOP. These alternative anbodiments have certain advantages and drawbacks as compared to the OFDM SOP. For example, the rnulti band SOP does not completely eiimmate ISI, but it is more robust to certain types of narrow band interfering signals because 4U the interference can be more confined within a given SOP bin as compared to the OFDM
SOP. The ISI that can be present in the mufti-band SOP could make it advantageous to use pre-equalization or post-equalization structures in conjunction with the spatial processing within a given SOP bin. While this complicates the spatial processing, the complexity drawback may be outweighed by other requirements such as robustness to irnerference or the 45 need to separate the SOP bins by relatively large frequency separation.
Only the OFDM SOP
will be analyzed in detail herein and it will be understood that one may exploit the other SOP
choices as needs dictate. .
As depicxed in Fig. 10, the ex~nplary SOP operates in the following fashion.
The symbols from the transmit spatial pre-processor, z~ (n) , considered to be in the frequency wo s~o93ss rcrmsrrnsi6o domain, are organized into Mr vectors of N complex symbo4s. Each of these vectors is then converted to the time-domaici using an N point inverse-fast Fourier-traasfoxm ~
procedure 205. Each of the parallel Mr time-domain sequences has a cyclic prefix added to the begirming, so that the last v demems in the IFFT output sequence form a pro-amble to 5 the N element IFFf output. The cyclic prefix operation is given by:
~z(1) ... z(~~ r ~"> ~s(N - v + 1) .. . z(~ Zil) ~ .. z(~~ r The application of the cyclic prefix is perforaa~d by cyclic prefix application procedure processor 207. For each antenna, the ( N + v length sequare passed to the RF
transmit chain for D/A corrversioa and modulation.
10 Ldcawise, each RF receiver chain produces a sampiai sequence of length N +
v .
Cyclic prefnc r emoval prxedure processors 206 remove the cyclic pns~x from each sequence by discarding the first v data symbols, resulting in MR vectors of N complex symbols. Each of these MR sequences is then processed with an N-poiirt fast-Fourier-transform ~.
These symbols are then passed to the receiver spatial processor.
15 The effect of the SOP is to substantially remove the ISI between any two symbols assigned to different bins, for any pair of transmit sad receive antennas.
Therefore, for each IFFT-FFT bin n, the recaaved signal values for each antenna, :(n) , aro related to the fitted frequency-domain symbols, z(n), through the expres~oa, =(n)=H(n~(n)+I(n) do (10) where ~c(n} is a complex Ma -dement vector at SOP bin n, z(n} is a compiac Mr .element symbol vector at bin n, and I(n) is the interference and noise at all receive antaonas for bin n.
Note that a time index is not included in the above equation since it is assumed that charnel is time-invariant over the length of a burst. The spatial sub-charnels, H(») , are MR by Mr element matrices that deSC~e the spatial comlation remaining in the wireless channel after the SOP. For the IMO case, each SOP bin may be charactaiud by a matrix of complex values, with each value represecrnng the path gain from a given tranarnit antenna element to a given receive elameat in that particular SOP bin.
To understarLd the result given by Equation(10), it is instructive to show how the SOP
pre-processor and post-processor acts upon the time-domain channel. The M>NIO
time-domain chancarl, G(k), contains M,QMr Toeplitz matrices that describe the time-domain input-output behavior of each antemia pair (see eq. I). This channel foraaulation is depicted in Fig. 10. It is wei! Irnown that by adding a cyclic prefix at the traas~mitter and subsequently removing the pre~c at the receiver, a Toeplitz intput-output matrix is transformed into a circulant input-outinrt matrix (the 'rth r ow is equal to the jth row cyclicly shiRed_ by i ; j elements). Therefore, the each G,,~ in Fig. 10 is transformed imo a circuiant G,,~ . The M>11~t0 cir~lant matrices are deliamted in Fig. 10 by G .
This particular class of SOP exploits the fact that any circulant matrix can be diagonalized by a predetermined matrix operator. One such operator is a matrix of the FFT
basis vectors. That is, for arty circulant matrix G_ , D = YGYx where D is some diagonal matrix and the scalar elements of Y are, 1 e-txmwrN , "~ - ,jN
Applying Mr IFFT opwations ax the traaamitter and MR FFT operations at the receive is described mathematically by a pro-multiplication of a NMR X NMR block diagonal FFT matrix wo ~os~ss rcrnJS~rnsi6o and post-multiplication of a NMr x NMr IFFT matrix. For example, the former matrix is due, Yu,t~

Therefore, inchrding the transmitter IFFT and receiver FFT operations, the input-output relationship is descnbed by a ... Dmr Yca..~GYu, - . . , where D,,~ is the diagonal matrix containing the SOP bin strengths for the antenna pair (i,~~.
Pre-multiplication and post-multiplication by permutation matrices Pr and PR
represents the collection of all antenna coons that cod to a common frequency or SOP bin.
This collection process, depicted in Fig. 10, results in a block diagonal matmc that relates the intguts and outputs:
H(1) 0 1'RYw,~GY~rI'r =
0 H(N) which is cquivaleut to Equation (10).
Spatial Prooasing The spatial prox~g pr~oc~re is now conaidaed. Since the SOP establishes N
MIMO spatial channels that are substantially independent from one another (Equation 10), one can consider the spatial processng within each bin separately.
Representative applicati~
of spatial pmc~aing to freque~y bin 1 wr~l be considemd as shown in Fig. 11 at the transmitter and Fig. 12 at the receiver. Fig. 11 shows M symbols: z( 1,1 ) through z( 1,11.
The notation z(n,m) refers to the symbol transmitted in bin n and spatial direction m. These M symbols will jointly occupy frequency bin 1. Each TSW 210A-C applies a weight vector to the symbol appearing at its input, and the dements of the resultant vector are routed to Mr summing junctions 21 I . One may consider the TSWs as being multipliers taking each input symbol and multiplying it by a vector that corresponds to a spatial direction in Mr -space. Furthermore, the M vectors define a subspace in Mr -space. Note that the TSW
vectors are considered to be column vectors in the discussion that follows.
When these M
vectors are collected into a matrix, the result is an input orthoganali~ng matrix or besleficisl weighting matrix for that bin. For each input bin, a vector including symbols allocated to subchannels corresponding to the bin is multiplied by the input orthogonalizing matrix to obtain a result vector, elements of the result vector corresponding to the various transmitter amenaa elements. Together, the TSWs 210A C make up one embodiment of a Tran~nit Spatial Processor (TSP) 230.
Each RSW 220A-C acceptsMR inputs, one from each receiver antenna path. Within the mw RSW, a weight vector is applied to the inputs (i.e. an imier-product is performed) thereby producing a received signal sample x(l,m):

wo ~3ss rcrms~rnsi6o x, (1) x(l,m)=u(l,m . , xut (1) where u(l,m) is the RSW for bin 1 and spatial direction m. Similar to a TSW, a RSW vector has an associated direction in MR -apace. Each RSW may also be considered to be a multiplier. This vcetor is considered to be a row vector. When these M RSW
vectors are collected into a matrix, the resuh is an output orthogonalizing matrix or bendcial weighting matrix for that bin. When a vector including symbols in a particular output bin produced by the SOP for each receiver antenna is multiplied by the output orthogonalizing the result is a vector including symbols received in that bin for various spatial directions.
Together, the RSWa 220A 220C feat one embodiment of a Receive Spatial Processor (RSP) 240.
Through proper choice of the weight vectors applied via the TSWs sad IRSWs, the M
spatial directions can be made substantially orthogonal to one another. The result is that the received signal sample x(l,m) depends only upon input symbols z(l,m) and not the M 1 other input symbols for SOP bin 1. Methods for selecting the TSP and RSP weight vectors are described in detail below.
The spatial pro<xssing desaibcd above can be applied to the other N-1 fiyuemy bins in addition to fiequency bin 1. The block diagram for such a system is depicted in Fig. 13 for the transmitter and Fig. 14 for the receiver. SOP processors 190 and 200 ensure that the frequ~cy bins remain substantially orthogonal to one while TSP 230 and RSP 240 that M substmot~ly orthogonal spatial channels exist within each frequency bin. The net result is that NM substantially parallel subchannels are constructed within the MIMO
communication system. In other words, the combination of SOP processors 190 and 200, TSP 230, and RSP 240 aeate a set of substantially independent space-frequency subchannels, x(n,m)=H(n,m)z(n,m)+I(»,m) dn,m.
This simultaneous aubstantisl orthogonalization of space and frequency can result in a scam increase in spaxrat efficiency since multiple data streams are being comawnicated thmugh the cham~el. Note that the mamba of substantially indepa~dent subchannels pos~le, in the multipath case, is equal to the number of SOP bins multiplied by the numbs of transmit antennas or the numbs of receive araennsa, whichever is stnaUer. Thd~cfore, the total number of spa~oo-frequency subchannels is less than or equal to N min ~li~l,.
,MR ~, when multipath is pre~t.
An exemplary set of TSWs and RSWs are derived from the value decomposition (SVD) of the spatial channel matrix for each bin, . H(n) = U(~)~(»)V(n)x .
The input singular matrix, V(rr) , contains Mr column vectors that define up to MT TSWs for bin n. Likewise, the output singular matrix, U(rr) , contain MR column vectors that when Hermitian transposed, define up to MR RSW row vectors for bin n. The TSWs and RSWs for other bins are de~rnined in the same fashion, through an SVD decomposition of the spatial matrix for that bin. Using this spatial processing, aubstantiauy independent multiple streams of symbols can be tran~itted and received. The strength of etch subchannel is equal to one of the elanents of the diagonal matrix E . These subcbannels strengths will vary.
Therefore, the subchannels will have varying signal to noise ratios and information capacity.
For this reason, it may be preferable to transmit and receive only on a subset of the possible subchatmels, or M < min ~41T ,MR ~ . For example, it may be improvident to use processing wo pc-r~rs~nsmo complexity on the weakest subchannels that may have a very small information carzying capacity. In this case, spatial processing is used to increase the received power of one or more parallel symbol strums. It may also be preferable to use coding tecimiques to leverage strong subchaanels to assist in the use of weaker subchannels. It may also be preferable to allocate either bits or tra~mit power among the subchannels to maximize the amount of information communicated.
The exemplary spatial processing described above requires cooperation between the transmitter and receiver to effectively orthogonalize the spatial channel for each bin.
Alternatively, this orthogonalization can be accomplished at only one end of the link. This can be advantageous when one end of the link can afford more computational complexity than the other end. In addition, spatial orthogona>iring at one end can be advantageous whm the channel model is known only at that end.
Consider the case where the orthogonalization is done at the receiver. Symbols are traasraitted along directions defined by some set of TSWs, v(n,m), When M TSWs corresponding to the same bin are collected into a matrix V(n), the composite spatial channel is, H'(n) = H(n)V(n).
This composite cham~d dthe MINiO channel in bin n from the M inputs to MR outputs. The spatial processing at the receiver can substantially orthogonalize this composite channel, H'(n), by applying appropriate RSWs even if the transmitter does no spatial processing. Let these RSWs be defined as the row vectors of the weighting matrix, WR (n) .
Two exemplary methods for determining WR (n) are referred to as the zero-forcing (ZF) solution and minimum-mean-square-error (M1VISE) solution. In the ZF
appmach, the weighting matrDC is the psedua-(left~imrerse of the composite channel, WR (n) = H'(n)l .
This results in, WR (n)H~(n) _ ~
where the identity matrix is M by M. Hence, the ZF solution, not only arthogonalizes the spatial channel for bin n, but it equalizes the strengths of each resulting subchannel.
HovvevQ, the signal-to-noise ratio for the various sub~unels can vary widely.
One sk~lod in the art will recognize that the ZF solution can result is amplification of the interference and noise unless the composite charnel, H'(n) , is early orthogonal to begin with.
An MMSE solution, on the other hand, does not amplify noise. For the MMSE
approach, t1~ weight, WR (n) , satis5es, .
min E~ WR (n)=(n) - z(n) ~~ ~, w"cA~
or, -, Wa (n) - Rsc~~c~~Ruw> >
where R=~"~ is the covariance matrix for u(n) and R=~"~~p~ is the crass-covariance between z(n) and =(n). Using, =(n) = H~(nh(n) + I~(n) and the fact that R,~~~ = a~s I ,results in the MMSE weight, WR (~r) = H'(n)x H'(n)H'(n)x + ~: Rl.
C

wo mss Pc rms~rnsi6o Note that when I(n) is spatially white noise, rhea R1= cr; I .
Similarly to the above orthogonalizstion at the reoaiver, the channel can be orthogonaliud at the transmit end wily. For this to ocxur, the transnritter is to have knowledge of the RSWs to be usai by the receiver. In a TDD duennd, where the channel achibits rcciproe'tty, these RSWs can be learned when that ever uses TSW directions equal to the ltSW
directions. Alternatively, the receiver may not do ar>y spatial procxssing, so the transmitter is responsible for spatial orthogonalization.
In this cax, the composite channel is 1 ~'(n) = U(n)$(n) where the matrix U(n) is composed of the RSW row vectors, u(n,m). This composite charnel describes the MlHtO channel in bin n from M,. inputs to M outputs.
Similar to the previous case, the transmitter can substantially orthogonaliee this composite channel, H'(n) , by applying appropriate TSWs. These TSWs are the column vectors of the weighting mabix, Wr (n) .
The traasmit weighting can be determined using the ZF or the MMSE approach. In the ZF approach, the weighbag matrix is equal to.the pseudo-(right~inverse of H'(n) . The MMSE solution satisfies _ min E~ H'(n)Wr (n~(n)+ I'(n) ' z(n) ~ ~~
wr<~?
Aa inrportaat simper to the garmal apaco-frequas3r Processing technique is the use of only one spatial ditaxion for each bin of the SOP. This case is depicted in Fig. 1 S f~ the transmitter sad Fig. 16 for the receiver. In this case, only N subchannels are cue. The N
input symbols, z(1,1) through z(N,1), are ptrocx:sed by N TSWs 210A B that weight and allocate these N symbols among the M? identical SOP processors 190. At the recesver, the atrtenna samples are proceasad by M,~ SOP processors 200. The MR SOP outputs corresponding to a common bin are v~ighted acrd combined in N RSWs 220A B.
With N
such waghtings, the result is N outputs, x(I,1) through x(N,1~ of the N
substantially ortlwgonal ~bchannelS.
When only one spatial direction is used in the TSP and RSP, one a~aplary choice for the particular weightings are the TSW and RSW direcxions that result in maximum subchannel strength. This maximizes the signet-to-noise ratio (SNR) of the signals, 3 5 x( 1,1 ) through x(N, I ). In this case, the opbmai was should satisfy, ~ ~ u(n)H(n)v(n) ~l, .c~>~c~~
with the implicit constraint on the sire (2-norm) of both the RSW weight u(n) acrd the TSW
weight v(n). To determine the solution to this optunization problem, consider the SOP
outputs for bin n when a single TSW, v(n, l), is used, . = H(n)v(nl)z(n,l) = h(nk(n,l) .
x,~~ (») wo ~s~o~ss rc~rms9~nsiso The quamity h(n) is refs to as the received channal vector. A chann~
idemific~tion technique is used to determine the received chancel vector.
Therefore, the optimal RSW weight is equal to the Hermitian of the received channel vector, h(n), u(n) = h(n)'~ .
5 Note that this is true regardless of the particular value of v(n). The optimal TSW direction, on the other hand, satisfies, maX u(n)H(n)v(n) = mar v(n)~' H(n)x H(n)v(n), ~c~> ~c~~
whore the optimal RSW direction has been used. The optimal TSW for bin n is equal to the scaled maximum eigenvector of the matrix H(n)x H(n) . One skilled in the art will recognize 10 that the optimal RSW is also equal to the scaled maximum odor of H(n)H(n)x .
A further advantageous simplification of the above techniques is the use of one or more common TSW and RSW directions for all bins. In other words, every bin has the same TSW and RSW weights. These weight vectors may also consider delay elements. In core embodiment, these weights are determined to maximize the SNR of the received signals, 15 averaged over frequency n. This is depicted in Fig. 17 for the transmitter and Fig. 18 for the receiver. Consider this embodim~rt with one spatial direction. In this case, the TSW and Rsw weig>us satisfy.
mar Ep~uH(n~~, 50 Note that the acpectation operator, E" , represents averaging over SOP bins.
This averaging 20 could also be done over amifiple bursts in addition to frequency. The solution to this problem is when v is equal to the maximum eigemrxtor of R, = E" ~H" (n)H(n)~, . 51 and a is equal to the maximum eigemrector of the covariance matrix formed firm averaging the outer product of the receive vector ~nnel, R~ = E" {h(n)h(n)" }. 52 The quay 'rny Rd is the spatial covariance maunc that de~ibes preferable directions to transmit to the desired receiver, a desired subspace.
This technique can be ga~aiized to the case where multiple directions are utilized.
In this case, MTSWs and MRSWs are determined to maximaze the average (ever bin) SNR.
received through the M spatial directions. The M spatial directions will not necessarily be orthogonal to each other. Therefore, there will be spatial crosstallc in the raxived symbols.
Multidimensional encoding and decoding techniques discussed below can then achieve a multipiic~ivt rate increase in the ice of such crossralk.
Alteroatively, the receiver can spatially orthogonaliu the subchannels by further wdgix:ing of the M outputs from the ltSWs. The composite spatial channel at bin n, with the RSWs and TSWs included is H'(n) = UH(n)V , 53 where the matrix U is made up rows equal to the RSW directions and V is a matrix with cohrmns equal to the TSW directions. Since U and V wero determined based on an average SNR criterion for all bins, the composite matrix H'(n) will not be diagonal.
Hence, the receiver can apply the additional weight, WR (n) , to orthogonalize H'(n) .
Alternatively, the transmitter can use the additional weight, W,. (n), to spatial orthogonalize the composite channel. Exemplary solutions for these weightings are the joint SVD, the ZF
and MMSE.
The advantage of this approach is that the processing r~rCd to adapt all N SOP
bin matzix WO 9809385 PGTlUS9Tl15160 channels may be substantially higher thaw the procxssing complexity to adapt the average TSP
and RSP.
The rejection and prevention of intce can be accomplished in conjunction witty the space-frequency processing discussed above. Thin is espeaally useful when the number of spatial directions used for communication is less than the number of antennas.
This case occurs when wean spatial directions are not utilised or when the number of antennas at the receiver and transmitter are not the same. In other case, one or both ends of the communication link have extra spatial degrees of freedom to use for the purpose of mitigating intce.
The amount of iuterfetence arriving at an antenna array can ba quantified by the imsrfennce covariance matrmt, R~ (n) = E~(nu(n)~ }.
where I (n) is the M,t leogt6 interference plus noise vector received in SOP
bin n. This matrix defines an undesired interference plus noise subspace in MR -space for bin n. The intce plus noise energy that contaminates a particular received subchannd symbol with bin a ate spatial directi~ m, is equal to, n x (n. m)R~ {n)u(n. ~) .
where u(n, rn) is the combining weight vector for the RSP{n,m). An advantageous interf~nce rejection technique is then to "whiten" the affect of the ince acxoss the spatid directions, so that the interference is minimized and spread evenly across all spatial directions used. T~~fore, each of the RSP weighting vectors are modifies by the matrix RJUZ(n).
a'(n,nt) - u(n,m)R~uz(n) .
Alternativdy, the RSP weighting vectors are the vectors of the output singular matrix, U'(n) , from tlas SVD of the modified spatial chemel, RI v z (n)H(n) = U.(n)~.(n)V.(n)x , Note that a very useful simplification of the above itttce rejection technique is to avaege the ititcovariance matrix ovtr all N bins and posat'bly a set of borate to strive at an avaagc spatial ice covariance matrix, R, , that is independent of bin n.
In this case, every RSP combining vector is modified in the same way due to latex.
This approach can significantly reduce the amount of computations needed to determine Rl and R~'m . Note that it is oRen beneficial to add a scaled matrix idtnaty term to estimates of the it~terl~crence covariance metciu to reduce the seasitnrity.of these intafmtnce mitigation approaches to covariance estimation errors.
Similar interfecrnce mitigation techniques can be advantageottsIy employed at the transmitter to reduce the amount of interference radiatod to unintentional receivers. In the TDD chatmd, reciproaty in the radio link allows the undesired receive interference subspace is each SOP bin to be ably used to describe the transmitter subspace. That is, the amount of interference transmitted to unintentional receivers is vx (n, m)RI {n)v(n, nr) , 60 where v(n, m) is the transmit weight vector for the TSW(a,m). An optimal iut reduction approach is then to minimize and '~rhiten" the transmitted intnce across spatial directions. In the same fashion as the receiver case, the TSW vectors are modified by ~ Rl a z (n) . Alternatively, the TSP weight vectors are the vectors of the input singular matrix V'(n) flrom the SVD of the modified spatial channel, wo 98109385 PCT/US9'7/15160 $~n)R J v= Vin) = U.tn)~.tn)V.~n~x . 61 Again, a significant simplification occurs when the into covariance matrvc is determined by averaging over frequency or SOP bins. It is especially advantageous to average over SOP bins in a fiequency-division-duplex {FDD) aystdn, where significant averaging of the receive corriariance matrix results in a good a of the transmit covariance, even though instantaneous channel reaprocity does not hold.
Irnerference rejoction at the receiver and interference reduction at the transmitter are done together by simply combining the two techniques outlined above. In this case, the RSP
vectors and TSP vectors are contained in the input and output matrices of the SVD o~
Ru=tn)HU)RrT=tn) As outlined previously, it can be advantageous to use the same TSWs and RSWs for all bins. This approach can be combined with intce lion by the d~rmining the transmit and receive weight vectors that maximize average power delivered to the receiver of int~st, while at the same time, minimizing power delivered to other undesirod receivers.
There are various optimization problems that can be posed to determine these TSP or RSP
diraxions, each involving the desired roceiver covariance matrix and the undesired covariance matrix. For example, one TSP problem is max vR Rdv such that vx R f v 5 Pt and v"v 5 PT .
That is, a TSP direction is chosen for all SOP bins that transmits the maximum amount of power to the desired receiver while maintaining a transmit power limit, P,. , and a transmitted interfera~e limit, P, . For this patticxrlar problem, the TSP direction is equal to the maacinwm generalized eigemrector of the matrix pair { Re, (R~ / P, + I / Pr ) ~. One example of an effective interference rejecting RSP for all SOP bins is a weighing that maximizes the average recaved S1NR. The RSP that maximizes SINK is the maximum eigemrector of the matrix R~»iReR~»:.
One fiuther simplification to the above algorithm is to model the interference and/or the desired covariance as diagonal, or nearly diagonal. Wl~ both R, and R~ are diagonal, the solution to the above optimization problem reduces to the maximal ratio SINK combiner, a, arui traasmittex, v. It is also sometimes preferable to only consider other subsets of the elemervts of either the desired or interference covariance matrices.
One skdIod in the art will also recognize that all the TSPs and RSPs eau be used when there is only one SOP bin, such as a common frequency.
Space F'rcquaicy Coding Mary of the advantageous apace frequency encoding techniques embodied in the invartion may be broadly classified in two eacemplary categories. The first category imrolves techniques wherein the spatial matrix channel within each SOP bin undergoes space frequency processing at the transmitter, or the receiver or both, resulting in a substantially independent set of one or more parallel communication subchannels within each SOP bin. The objective of the encoder and decoder in this case is to appropriately allocate the transmitted information among multiple independent space-frequency subchaaneis using interleaving, power and bit loading, or both. The second category of space frequency encoding involves transmitting and receiving one or more symbol sequences in each SOP bin using one or more transmitter and receiver weight vector combinations that are not necessarily intended to create independent spatial subchaonels within each SOP bin. This results in significant cross-tally between each symbol stream present at the r~eceiv~r output. A decoder then uses knowledge of the wo ~3ss rcrmss~rnsi6o equivalent mat<ix channel within each SOP bin, and knowledge of the set of possible ~coder sequences to estimate the encoder symbol sequence that gave rise to the cross-talk rich output SOP bin vector sequence. The main differentiating feature bewoen the first approach and the second approach is the presence or lack of spatial processing that results in substantially orthogonal spatial subchannels within each SOP bin. Both approaches have the advantageous ability to multiply the data rate that can be achieved in M1M0 channels with muitipath.
Coding for Ssbswatiaity Orthogonal Sprco-P'r~qneney Subc6anadt In applications where the spatial channels are processed to achieve muhiple substamially independent spatial subchannds within each SOP bin, an advantageous embodiment of the im~tion involves coding the input data sequence into a digital symbol stream that is then routed in various benefiaal ways through the available parallel space frequarcy aubchnnnc(a. Fig. ZZ depicts a preferred embodim~t. This embodiment invol es distributing the symbol outputs of a single among all of the available space frequency aubchnnnc(a. Several known coding schemes that can be combined effectively with space frequency procxssing to distribute information tranon over the spacx and frequency dimensions of a communication channel. This discussion assumes estimation of the MIMO
channel by transmitting a series of tra>ning symbol sequences from each antenna element as disaiased heaein. Tlu disa~sion &~rth~ assumes that the receiver and transmitter share the information required to decompose the channel into parallel sub-channels, or the TDD toclnnques discussed herein are use to do the same.
Referring again to Fig. 22, the pembodiment exploits a three Isyer coding system. The first layer of coding includes the combination of transmitter TSWs through 210B, Transmitter SOP processors I90, receiver SOP procasocs 200, and receiver RSWs 220A through 220B. This first layer of coding performs the spatial processing. The second layer of coding includes a Trellis Encoder and lnterleaver (420) at the transmitter in combination with a De~ntaleaver and ML Detector 430 at the receiver. The third layer code imrolves Reed Solomon (RS) Encoder 410 at the transmitter in combination with as RS
Decoder 440 at the rte. The bit level RS coding occurs prior to the trellis encoding and the Rted Solomon codeword daxtor acts upon the bit sequemx from the ML
detaxor. The fourth, layer of coding involves an ARQ code that recognizes Reed Solomon codaword errors at the receiver in the Receiver ARQ Buffer Cormol 450 and requests a codeword retransmission from the Transmitter ARQ Buffer Cornrol 400. Ttas retransmission request is made through a Reverse Link Control Channel 460. The reverso control channel is a well known radio system concept and will not be discussed herein. This combination of coding techniques and space frequency processing is preferable because it provides far a rich combination of space and frequency diversity and it is capable of obtaining very low bit error rates. The detailed operation of tlm RS encoder and decoder, as well as the ARQ system is well known to one skilled in the art. Following this discussion, it will be clear to one skilled in the art that other combinations of one or more of these four coding elements may be anployed with advantageous results in various applications.
The trellis coding step may be substituted with CBM.-QAM or a turbo code.
Similarly, tlu Raed-Solomon code may be substituted with a block code, or with an error checking code such as a CRC code. The transmitter end would then include the necessary encoder and the receiver end would include the necessary decoder.
There are at least two basic methods for employing trellis coding to distribute information among substantially independent space frequency sulxhannels. One method is adaptive encoding that modifies the bit and power loading for each subehanel according to its wo ~3ss rcrws9~nsi6o quality. The second method involves maintaining constant power and bit loading for all space frequency subchannds. Both of these methods are discussed blow.
Space Frrqnency Trellis Coding with Orthogonal Spatial Sufxhannds and Adaptive Power and Bit Loading Fig. 22 depicts the coding and interleaving detail for the transmitter and receiver portions of the present embodiment. Encoding and Interleaving system 10 encodes the data into a set of complex symbols. Each of the complex symbols is then allocated to a particular transmitter TSW (210A through 210B). The input to each TSW foams a vector of frequency domain symbols that are fed into the same bin of one or more transmit SOP
processors. Each transmitter TSW, poss~'bly in conjunction with a receiver RSW converts the matrix channel within each SOP bin imo a set of substantially orthogonal space frequency subchannels using one of the methods discussed herein.
Fig. 23 displays a more detailed diagram of the encoder sad intedeaver. An Information Allocation Unit 360 assigns the bits and the transmitter power that will be allocated to each space frequency subchanmel. One method for accomplishing this assignment is the so-called gap analysis. In this technique, a particular coset code with an associated lattice structure is characterized by first detaining the SNR required to achieve a theoretical capacity equal to the de~red data rate. The code gap is then the SNR
multiplier required to achieve the target probability of error at the desired data rate. In a parallel channel communication this gap can be used to determine the power and bit distributions that maximize data rate subject to a probability of error constraint. wth a coding gap of a, the rate maximizing water-fiiling solution for the space frequency subchatrrrds becomes z p(n, m) = ~ - °~" a i , ,~,(n, m~
where Q"Z is the noise power and m is the spatial index and n is the DFT
frequency index.
The bit allocaxion per sub-channel is then given by b(n,m)=1 1+p(n'm)~~'sn'm~
aQ"
After the power and bit loading assigmnents are accomplished in the Information Allocation Unit, the bits arc encoded with a Trellis Encoder 370. It is not possible to achieve infinite bit resolution (granularity) with coast codes. Therefore the gap analysis solution should be modified. Several bit loading algorithms exist to resolve this problem. One method involves rounding down the water filling sohrtion to the nearest available quantization. The gr~arrularity of possible bit allocations is determined by the dimensionality of the coast code lattice structure. In the MIMO channel communication structures described herein, the orthogond constellation dimensions are the complex plane, space, and frequency.
Fig. 26 illustrates an example of a practical method for bit loading with a trellis encoder that uses a one dimensional QAM symbol constellation. The bit toad is adjusted down for a given trellis encoder output symbol by assigning a number of fixed zeros to one or more of the input bits to the encoder. Fig. 26 shows the operation of the trellis encoder and the trellis state diagram for the decoder for four successive symbol transmissions. Four bits are assigned to the first space.frequency subchannd. A first subchannel trellis encoder input 350 is assigned 4 bits so Symbol 1 can take on any one of 32 values. There are two bits feeding the comrolutional encoder, and two bits feeding the coast select. The ML detector at wo rc~rros9~nsi6o the receiver uses the trellis state diagram and the chanaei state information to solve the maximum likdihood recursion. This is efficiently accomplished with the Viterbi algorithm.
The trellis code state diagram defines a set of symbol sequence possibilities {Z~. The space w frequency subchannd is denote H(n, k) , for SOP bin n at burst k. The maximum likelihood 3 equation is then given by (I)r~s(2)r~...~z(~T~=~=s~~~~',=cz :..,scX~r~~~H(n'k)z(n) ~n'k~~~
where z(n) is the symbol hypothesized for SOP bin n. The decodes output state diagram for frost space frequency subchaand 340 includes four possible parallel transitions for each trdlis branch and atl of the trdlis branches are poss~'ble. The second space frequency subchannd in 10 the sequence is assigned three b'tts so a second trellis encoder input 352 shows one bit fed into the coset select with two bits still feeding the convolutional encoder. A
decoder state diagram 342 for the second space frequency subchannel has only two paralid tr~ansifions for each trellis branch but stilt maintains all trellis branch possibilities.
Continuing is succession, a third space frequency subchannel is assigned only two bits to an encoder input 354 so there I S are no parallel transitions considered by a trellis decoder state diagram 344. In a fourth space frequency subchaaad, only one bit is assigned to an encoder input 356 so there are no parallel transitions and some of the trellis state branches (346) are no longer considered by the decoder. It is undustood that FIg. 26 is provided as a graphical aid sad is not intended to represent an acxual desigQ
20 The maximum Euclidean distance error sequence design metric is one pre6a~aWe choice for a trellis encoder used with the parallel space frequency channel with this bit and power loading embodimern of the imrornion. Other code error sequence design metrics that are advarnageous in various application conditions include product distance and periodic product distance.
25 Referring again to Fig. 23, the output of the encoder is interleaved across the various space frequency suixhaanels using Interleaving block 260. Typically, the interleaving process distributes the symbols so that symbols that are near one another at the encoder output are well separated in both the SOP bin assignment and the spatial subchaimd assignmeut. This distn~utes the effects of channd nation errors and localized frequency domain or spatial domain interference so that the decoder error is reduced. It is understood that the bit and power assigooxnta by Information Allocation block 360 take piece with lrnowledge of the post-interleaved chesnel strength. It is understood that the encoding and decoding process can begin and end within one burst, or it may take place over a multitude of bursts.
One skilled in the art will recognize that a multitude of less soghisticated adaptive power and bit loading algorithms can be advantageously applied to a substantially independent set of space frequency subchannels. One example is an algorithm wherein a space frequency subchannel is either loaded with maximum power or no power and the bit distribution may be adjusted in only two increments.
A second alternative embodiment shown in Fig. 19 includes one encoder for each SOP bin, with the output symbols of each encoder allocated among several spatiat subchanaels. A third embodiment shown in Fig. 20 imrolves one encoder for each spatial subchannel, with the output symbols of each encoder distributed among the SOP
bins for that spatial subcharnml. A fourth embodiment shown in Fig. 21 irnohres a separate encoder for each available space frequency subch~el.
It will be clear to one skilled in the art that the chaat~el estimation tools taught herein are very useful in itiqaroving the acauacy of the channel estimates used for the bit loading and decoding process.

wo Pcrnrs~nsi6o One skilled in the art will recognize that many of the other coding techniques for parallel sub-channel bit loading comanuucation systems, not mentioned lu=re, can also be applied to the gresent invention.
Space Fr~equenncy Trellis Coding with Orthogonal Spatial Snbchannela sad Flat Powex and Bit Distribndon In some cases it is di~ta~lt to adaptively load the power and bit assignments for each available space frequency subchannel. For example, the transmitter and receiver may not be able to adapt the loading fast enough to accormmodate time domain variation in the channel.
In another e~le, the required feedback from the receiver to the transnritter requires a significant portion of the available reverse link bit rate. Adaptive bit loading may also be overly complicated for applications. Thus, it is often advantageous to encode and decode a symbol stream in such a mamier that the power and bit allocation is constant for all space frequency subchannels. This is easily accomplished by employing the embodiments depicted in Figs. 22-23 , and assigning a constant power earl bit allocation to all space fi~equency subchamels in the Information Allocation block 360.
Spaces Frequency Coding Witlwut Orthogonal Spatial Subchanne4 In applications where tlu spatial chaonds are not processed to achieve substantially orthogonal spatial aubcbannels within each SOP bin, an advantageous embodima~t of the inve~ion involves utilization of a vector maximum likelihood in the receiver to decode a symbol sequence that includes multiple ayrnbols per SOP bin. The vector maximum likelihood detector is capably of determroing the tzans~aittad symbol vector_in each SOP bin even in the presence of spatial subchannels that contain significant cross-coupling between the channels. The vector maximum likelihood detector uses an estimate of the matrix channel from each SOP bin to da;ode a sequence of of symbols with one group for each SOP
bin. The groupings will be referred to here as a multidima~sional symbol vector, or simply a symbol vector. The ML detector uses an estimate of the matrix channel that exists in each SOP bin to fmd ttas most likely sequence of transmitted vector symbols.
Wig. 24 depicts a transncritter system wherein multiple spsce/frequency subchaimela are eatployed without spatial orthogormlization. Fig. 25 depicts a receiver system for this appGcxition.
The bit sequence b(k) is encoded irno a sequence of multidimensional symbol vectors in a Bit to Symbol Facoding block 250. Each output of the encoder is an Mo by 1 compi~
symbol vector, where Mo is the mmtba of spatial dira~ions that will be used for transmission.
Note that Mo is preferably chosen to be Less thaw or equal to Mr. A preferable construction of the encoder is a multidimensional trellis encoder. One advantageous metric for designing the trellis ~oder conion and convohrtional ~coder polynomial will be provided below. Within the praviousIy discussed Symbol Interleaver block 260, the vector symbol sequence is demultipiexed and interleaved with a Symbol Sequazce Demultiplacor 300 and a Transmit Symbol Routing block 310. Transmit Symbol Routing block 310 irneslesves the vector symbol sequence so that the elements of a given vector symbol are grouped together and transmitted in one SOP bin. Thus, different vectors are separated by a multitude of SOP
bins before fission, but all d~eats within the vector symbol share the same SOP bin.
The purpose of the interIeaver is to distn'bute the vector symbol sequence so that the fading prexent in the nwithin the SOP bins is r~andomi~ed at the output of the receives interleaver. The decoder can recover information associated with cymbals that are wo ~3ss rcrrtrs~nsi6o transmitted through SOP bins that experiarce a deep fade, provided that the adjacent symbols do not also experiaice the same fade. Since there is often a high degree of correlation in the fading experiencxd by adjacent SOP bins, the interieaver makes the fading more random and improves decoder error performance. After interleaving, each element of a vector symbol is assigned to one antenna for the SOP bin assigned to that vector symbol.
Transmitter SOP
processors 190 perform the ttansmittar portion of the SOP.
After the transmitter SOP, it is often advantageous to perform spatial processing with TSP 230. It is understood that the matrix representing the operation of TSP
230, i.e., the Traasmitta Weight Mat:ix, may also be an identity matrix so that no weighting is implemented. It can be bena$aal to choose a number of spatial directions that is less than the number of tnmsmitter anti. In this case, the Tran~nitter Weight Matrix inc~ases the dimensionality of the time domain vector sequence from the SOP bank As an example of when it is advantageous to choose a subset of the available transmitter spatial directions, if the receiver has fewtr antennas than the trarram'rtter, then it is known that the information capacity of the matrix channels within cacti SOP bin will not support a numbs of parallel information subchannets that is greater than the number of receive antennas.
This implies that the number of symbols in each tran~itted symbol vector, and hence the number of transmitted spatial directions, should not be greater than the rnrrnber of receiver antennas. As another example, in a Rayleigh fading channel, the smallest singular values of an MR by Mr 2U matrix channel are on average much weaker thaw the largest singular value.
This implies that the average information capacity contained in the smallest siagu<ar value may not justify the extra signal processing complexity required to transmit over that dimension.
Ia both of these cases, it is advisable to choose an advantageous subset of the available transmit spatial directions.
The otter may not have knowledge of the individual channel matrices within each SOP bin but may have knowledge of the covariance statistics of the charmel matrices, averaged over frequency, or time, or both. In such cases, the Transmitter Weight Matrix can be optimized to select one or more spatial drections that maodmize the average received power for the chosen raunber of spatial directions. The procedure for optimizing the Transmitter Weight Matrnr for this criteria is def~d by Equations 50 to 52 and the associated discusaon. This is one preyed method of selecting an advantageous set of spatial directions for the Transmitter Weight Matrix. Another advantageous criteria for selecting the transmitter spatial dira~ions is to maximize average raxived power subject to constraints on the average into power radiated to uninteutional rccxivers. The procedure for optimizing the Transmitter Weight matrix for this criterion is definal by Equations 60 and 61 and the associated discussion.
After the time domain signal is ~atiallY Processed, the signal is upconverted to tfm RF
c$rrier frequency using Modulation and 1tF System 40 before being radiated by Transmit Antennas 51. Referring now to Fig. 25, at the receiver the signal is downcomrerted and digitized by Receive Antennas 111 and Demodulation and RF System blocks 120.
The RSP
240 may then be used to process the time domain signal. The operation of RSP
240 may be characterized by a Receiver Weight Matrix which may be an identity matrix. One emlmdiment irrvolves optimizing the RSP weights to reduce the number of received signals from MR to Mo, which is the number of elements in the transtnitted symbol vector and is also the number of fitted spatial directions. In this case, the Receiver Weight Matrix can be optimized to increase the average signal power in each received spatial direction. The opmnization lure to accomplish this is defined by Equations SO to 53 and the associated i wo saro93ss rcrnnsmnsiso Channel ID block 130 is used to ~dmate the matrix channel in each SOP bin.
Procedures far channel estimation are descn'bed below. Channel state information for each SOP bin: is fed into a Symbol to Bit Detecxor 280 which decodes the symbol sequence after it is passed through a Symbol Deinterleaver 270.
S At the receiver, after de-interleaving the SOP bins, the space-frequency sequencx is again comrat~ into a serial symbol stream by Demultiplexor 300. For a graven set of spatio-w temporal vector symbol sequence possibilities {Z}, and an estimate, H(n, k) , of the channel matrix in each SOP bin n at burst k, the maximum likelihood detector is given by equation (70):
{z(1)r,i(2)',...~Z(~T~=arg~~~~~~'.~c--~cN~'~~~R'(n'k) S~H(n'kh(n) =(n'k)~x~
where t(n) is the vector representing the code segment hypothesized for SOP
bin n, and Rr(n,kj is the estimated noise plus interference covariance rnarivc for SOP
bin n and time k.
This equation can be solved eff ciently using a vector ML detector. The SOP
bin cheumel matriac estimates are understood to include the effects of the Transmitter Weight Matrix and 1 S the Receiver Weight Matrix. It is uiuierstood that the noise pre-whiteaia~g step in the ML
detxtor cost ftmcdon can be substituted by a bank of RSPs that pctForm the interferatce whitening as descn'bed herein.
In a Rayleigh fading channel, a desirable manic for dig the trellis code is given by the product of a sum involving the two-norm of vector is of the trellis code error sequence:
hl~~n~=
where q is the mrmb~ of SOP bins in the error sequencx, and e(nj is the vector differ~e baween tlse tW a mufti-dimensional code symbol stg~nnent and the incorrect mufti-dimensional symbol code aegmern for SOP bin n. This code design metric is a generalization on the coonal product distance metric which contains a scalar error entry in the product equation while the new code design metric contains a vector two norm entry in the product equation. It should now be evident that the multidimensional encoder can be realized by either directly produang a vector consi~ng of a multidimensional QAM symbol with the encoder output or by grouping complex QAM symbols from a one dimensional encoder output into a vector. 'The vector symbol encoder alternative is preferred in some cases because this approach provides for a larger metric search result and hence a better fading code. After deinterleaving, the decoder that is used at the receives searches over all possible multidimensional symbols within each SOP bin to ma»dmize Equation 70. It is understood that one skilled in the art will recognize after this discussion that other desirable metrics such as Euclidean distance metrics, metrics designed for Rician fading channels, periodic product distarnx metrics, and others are straightforward to construct and space-frequency codes can then be determined through well known exhaustive search techniques.
In either the one dimensional encoder case, or the multidimensional encoder case, the encoder constellation selection and code polynomial search to maximize the metric can be carzied out using a number of well known procedures.
It is possible to improve the performance of the space-frequency coding system died above by-usiag a number of transmitter antern~as, or a munber of receiver antennas, that is greater than the number of symbols transmitted in each SOP bin. If the number of receiver antennas is greater than the number of symbols in each SOP bin, then simply applying 4S the approach descxibod above is advantageous. If the number of otter antennas is WO 98/09385 PGTIfJS9'1115160 than the nu~a of symbols transmitted in each SOP bin, thm the te~ques embodied in Equation 70 are advantageous.
Channel Id~ntificstion The opeistion of Channel Ide~cation block and Training Symbol Injection block will now be descn'bed. The transceivar should determine the MIlNiO chaimel in order to form the TSWs and ltSWs. For coherent spatial processing and detection, the receiver should obtain an elate of the chazmd. We wish to ide~r the set of matrix chancels that results aBer processing by the transmitter and receiver portions of an SOP. The notation for this channel is H(n) d n wha~e n is the SOP bin index. C6anod identification techniques embodied herein can be applied to several preferable SOP pairs including the IFFT-FFT with cyclic preinc, the nwloband filter bank, or auy other of a numbs of well-known SOPS. The following exemplary chancel identification approach exploits the coitdated frequency fading across and poas~bly the coed time fading in the channel. The correlation in the frequency domain arises due to the limited time delay spread of the multipath channel.
The correlation in tame is due to the fact that the channel, while time-varying, is driven by bead-Limited Doppler frequencies created by objcets, which can include the transmitter and/or receiver, moving in the physical environnarnt.
The vvirdess link is bidirectional, tha~afore each end of the link should ete not only a receive channel, but also a transmit channel. For example, a base station should a both as uplink and downlink channel. In systems which employ time division duplexing (TDD), electromagnetic recdproaty implies the receive and transit propagation enviromnents are the same, the transmit channel to be estimated from the raxive channel. I3owever, the transmit and receive electronic responses are not necessarily reciprocal, amt because the net channel response includes the electronics, a calibration procedure should be used to account for these differences. This calibration procedure provides for matching in the amplitude and phase response between the multiple transmitter and receiver fi equency converters. Several TDD calibration procedures are known in the prior art and will not be discussed herein.
In systans employing frequencyr division duplexing {FDD~ the propagation meditma is not reaprocal; however, the paths' angles and average strengt6a are the same for transmit anti reeve. This enables the lice of subspace reaproc~r, but inwra a more ri~mus calibration requir~aeut. The FDD calibration should insure subspace reciprocity which requires that the array response vector at a given angle on recave is proportional to the corresponding vector on transmit. This requirement is satisfied by again calibrating the amplitude and phase diff~ances among the multiple transmit and rccdve frequency comert~ chamds and by matching the transmit and receive antenna elemem response as well as the array geometry.
An alternative approach to transmit channel estimation in FDD sy~ems uses feedback.
The transmit channel is Bred by sending training symbols to the receiver, which records the amplitude induced by the training symbols. Using receiver to transmitter feedback on a separate feedback comrol channel, the training responses are sit back to the transmitter. The transmitter, knowing the training excitations it used and the corresponding responses through feeedback, the forward cham~d can be estimated.
In general, channel identification can be done either with or without training. A
desrable channel identification algorithm should be robust to and operate is a variety of modom implementations. A preferable Mll~iO channel identification technique operates with embedded training inserted into the data stream by Traiming Symbol Inyection block 20 in each burst. In this case, both data symbols and training symbols may be transmitted within a single burst. Furthermore, the channel can be determined in one burst, or f ltering training wo ~o~s ~ rcrrtrs~rnsi6o data gathexed over multiple bursts. Bang able to update the channel estimates after every received burst makes the overall communication system robust to time variation in the charnel. In addition, frequent channel estimates reduce the destructive of imperfect carrier frequency recovery. Since imperfect carrier recovery imparts a phase shift to the 5 channel that continues to ,grow with time, shortening the time between charnel estimation events keeps the channel ion information from becoming "stale". Note, howaver , any of the well known blind channel estimation techniques can be used to determine the training symbol outputs as an alternative to using training. However, adaptive blind training is more prone to gech~ng burst snore.
10 The parameters to be identified are the N MI11rI0 spatial channel matrices.
Hence, there are N ~ MR ~ Mr complex elements to be deteravned, H~.i~n~~ do E ~1,NI , 'di E ~l,Mx I , brj ~ I~Mr I~
By exploiting whatever correlation exists across the SOP bins, it may be possible to reduce the amount of overhead required to identify the channel. The amount of correlation that 15 exists across SOP bins is determined by the specific implementation of the SOP. If the SOP
implementation includes the IFFT-FFT pair, and the length of the FIR channel is time limited with v « N, then a relativellr large degree of correlation exists across the SOP bins.
In certain embodiments of the invention, the desired technique should identify the MIMO channel on a burst-by burst basis, such as those with rapidly time-varying channels.
20 This implies that training data should be included in every burst. If the throughput of information is to be maximised, the amount of training data in each burst should be minimized. It is ther~ore useful to determine the minimum amount of training data required, per burst, that aDows full characterization of the channel by the receiver. It turns out that the minimum mimbtt of training symbols required to sufficiently excite the MW IO
chamiel for 25 G~timation with an OFDM SOP is Mr v . To understand this result, consider the identification of a SISO channel, where each of the N values of the vector H,,~ should be found. These N
values are not indep~ident since, $~.~'l' 0 °X~'Z, where X is a vector of all SOP bin outputs for antaina i , Z is a vector of bin inputs for 30 antemia j, and h is a vector of the time-domain FIR channel from antenna j to antenna i .
The matrix operator ~-' represents element by element divide. Since the time-domain chanr>cl is time limited to v samples, only v values of tlm transmitted syrr>bds, Z , need to be training values. Furthermore, the identification of the SIMO channel only requires the same set of v transn>itted training tones, since each SISO component in the SIMO channel is excited by the same input data. In a system embodiment with multiple inputs (Mr > 1~, identification of the MII~fO channel requires the idezirificataon of Mr separate S1M0 channels. Hence, only Mr v training symbols are needed to sufficiently excite the MQvIO
charuiel for channel identification.
MB~IO ident~tion The iderarfication of the MIMO channel is accomplished by separately siting each of the transmit antennas that will be used for communication. This da;omposes the MIMO
identification problem into Mr SIMO identification problems. In order to accomplish cl>anrid idon in a single burst, Mr nnmially ~ sets of v bins are selected final the N available bins to carry training symbols. Each transmitter antenna carries training W~ 98/09383 PCT/U897/15160 symbols in a unique one of the MT sets of bins, while ttanamitting no energy in the bins contained in the union of the remaining Mr -1 sets of v bins. Thin is accomplished by choosing the TSWs 210A-C that correspond to training bans such that a single army in the vector is "1" and the remaiinng entries equal to "0". It is the,"' entry of a TSW that is set equal to "1" for those training symbols which an to be tran~tted from the f a antd~a. For example, say that symbol bin n = 2 is one of the training bins associated with ttan~nit antenna 3. Tlxn, TSW(2,1) s ~0 0 I 0 ~ ~~ Of , and TSW(2,m) = 0 for dm ~ 1, and the corresponding uainiag symbol z(2, I ). BY examining the cxmtecrts of each set of tra>niag bins separately, the M»IO channel response is determined by finding MT
independem SIMO charnel responses.
In embodies in which rapid updates of the cbannd estimate is not required, another exemplary training scheme may be employod. This training scheme imrolves using just one set of v training bins. On a given lnrrst, one of the transmit antamas sends training I 5 symbols in the training bins and the other antennas transmit no energy in those bins. This allows the receiver to identify one of the Mr SIMO . On the next bury, a different antenna sends training symbols in the training bins while the other aatem~as transmit no enargy in those same bins. The receiver is then able to identify another set ofN SIMO
channds. This procedure is repeated until tr'simng data has barn seat by each of the transmit antemms, Blowing the entire IvBMiO channd to be identified. The entire pro~dtu~e is repeated cxuninuousiy so that full channel is deterraidod ~ Mr bursts.
SIMO chanad idenon We have just shown that identificstian of the MIMO channel can be accomplished by successive identification of each S»NIO cbaanel. It is therefore usef;rl to discuss specific tech for obtaining a SIMO chenad response. The following discussion assumes that the SOP is the IFFT FFT pair. C>umad ideatif~ion techniques for other SOPa that exploit frequency and poasbly time cotrdation in the sitar &ahion will be obvious to one s)cillod in the art.
It is assumed that a certain subset of available SOP bins are allocated for training. Let J be this set of frequency bias used for learning a S>ZVIO channel. To begin, assume that J
contains v bin indicts. Furthermore, let Zr be the v training symbols and X«
be the received data in the training freque~ncy-bins from arrears i. Let the quantities h, , H, be the estimated timo-domain and frequency-domain channels from the transmit antenna under consideration to the recave antenna i. In other words, Ir, is the v-length impulse response from the input wader conaidmation to output i. L~cewise, H, is a nectar of N
frequency domain values for this channel_ With these de0nitions, it can be shown that (gl) and where, Hr =Yp.~t ~ ( v = ~1,2,..., v}, wo ~o~ss rcnrtrsrrnsiso N = ~1,2,...,N}, and yrQ = ~Q ~~ro~N ~ dP E ~p~ ~9 a ~~ .
This also generalizes to any number of training tones, r , in which case the set J includes r bin indices. When y z v, the frequency domain channel can be determined by, Hi = Yp.UY~ v'Yr.v~~Y v(Xr.~ ~' Zr)~ t83) Note that mauy of the above calculations can be performed in advenx if the bins are predetermined and fixed. Then, the matrix Y~,,~ (Y TY~~ r' Y r can be computed and stored.
Note that there is no requirement that the training symbols always reside in the same bins from burst to burst. As long as the transmitter and receiver both know where the sya~ols are Placed in any given Must, the trairdng bins may be varied from one burst to the next. This may be useful to characterize the nature of colored (across SOP bins) noise and/or interference are prat.
A highly advantageous simplification of (83) can be done when v training symbols are phuxd in bins that era evenly spaced throughout the burst. In other words, J = ~0, ~,~,»..,~~. In this case, Y~ ~ is equal to the v-pouit IFFT matrix so that w equation (81) represanS the tion of an v-point 1FFT. One may then obtain H, in w equation (82) by performing an N-point FFT on a vector consisting of h, padded with N - v zeros. This approach to id~fying H, is only of ion order (N + v)log~ v .
Identification over multiple bunts Idet~f c~tion accura~5r cau be improved by increasing the of tta~g symbols within each burst or by averag;ng over multiple bursts if the channel is correlated from one burst to another. Some degree of time donain correlation exists in the channel bee the Doppler frequ~y shifts caused by mrnring objects in the physical erwironmeat are band-limited. This time correlation can be exploited by recursively filtering the estimated channel from the present burst with channel est>mates from previous bursts. A
general $ltering approach is repr~ted by h(k + 1) = F(k)h(k) +G(k)h(k) w>tere 6 is the Smoothed channel estimate of h over bursts k. The psrticarlar recursive filter ~ghts F(k) arad G(k) c;~un be derived in a of faslrions. Two exemplary ~t~ing methods are given in the following. The first approach detercnin~ a timo-imrariant FIR filter for each element of h based on a MMSE cost function. The second design is time-varying Ka>msn filter.
A particularly simple, yet effective, filter design technique is the determination of a time-invariant FIR filter, w, that minimizes the MMSE between the true channel impulse response and the filtered e.~timate. This design approach is referred to as Wiener filtering. In this embodiment, independent fading is assumed on each element of the channel impulse response. Therefore, each element of h can be considered independently. An FIR
filter produces a f ~wtimate by forming a weighted sum of the previous p+1 estimates for that particular impulse response element, wo s rcrrt~ss~nsiso w h, (k) k = ~"
h, ( ) . 'd~ =1,2,..., v .
~tk-P) Using v such identical filters for catch element of the impulse response, then the frltered estimate is given by h(k) _ ~lr, (k) ~ ~ ~ h" (k}~ . The Wiener filter solution for w satisfies the following equation, w Z
h,(k) miaF.'~Iwxlt,(k)-h,(k~=)=m~'t~nE wx w . -hyk) hr(k-P) The solution for the above optimization problem is givoa by, 'e'=~y(k~~(k)x~'~r(k~r(k)1=R., 'R.,w .
If each delay in the channel impulse response undagoea Raleigh fading is assumed then, JoU~ Jo(m(IY+ v)T) ... Jo(m(N+ v)T) Jo(m(N + v~') ' . . z R" - Q~ , +a,I
JoUP(N+v)T) ... ~~0(0) and Jo~~
z Jo~m(1V+v)T~
8,1,x, = cn , Jo(ap(N+v)T~
where T is the sampling rate, m is the maxiurum Doppler froquency, and Jo is the zeroth-order Bessd function. The quantitiesQ; aril e; are the average chard power and the channel estimation posse power, respecdvdy.
This filtering approach has many advantages. First, it is computationally ample. Each coe~cxent of the channel impulse response is f ltered independaitIy with a constant, precomputed FIR weighting. Second, the underlying time-correlation in the multipath fading clutrmd is effiaently exploited. Third, the exact values used for the f iter are optimal in a MMSE sense.
A more gmer~iz~ time-varying Sltecing approach is now developed based on the Kalmaa filtering equations. A general model for ttm timo-cowdated nature of the channel impulse response is given by the following set of equations, f(k+1) = Af(k)+q(k) 6(k) = Cf(k) +r(k) where the q and r represent noises with covariances Q and R, respectively. The matrices A,C,Q,R are used to define the particular model for the correlation of the impulse response wo ~sro~s rc~rros~nsiso over bursts. Note that the vector h can also include the impluse response coefficients for more than one receive arneana. In this case, the above model can include both time correlation and correlation across space.
In a mufti-access scheme, successive channel identifications may occur at an irregular rate. In this case, this Kalman filter approach is particularly useful since the filtering can be done with measurement updates and time updates, f (k + 1) = A(I - L(k~)f (k) + AL(k)h(k) L(k) = P{k~~' (CP{k~" + Rr' p{k) = A(I - L(k~)1'(k)A H + Q
h(k)= Cf(k) where L(k) = Owhen the receiver is not receiving data in the present burst.
Interference Sub:tuce Identification For many of the spatial processing techniques embodied in this invention, the opa~ation of the TSP and ItSP can depend, in put, on the Ievel of intaf~a~ce present in the t 5 wirdess enviromnent w~ which the iron operates. More specifically, it may be preferable to reduce the amount of interference comributed to other receivers by a judicious choice of the TSWs. It may also be preferable to improve the signal quality at the receiver by using ltSWs that reject interference. In these cases, some quantitative measure ofthe interference across apace and frequency is needed.
One preferable meas~e of the interference present is the so-called interference spatial covariance matrix, which describes interference cmrelabon across space for each frequency ~4 Run)=~;(n)I(n)H~. (1) where :~ (n) repres~ts an MR -length racW rod vector of signals from the int~ng transmitter(s). To be more precise, R, (n) descn'bes the interference and noise correlation across space for each frequency bin. Since we assume that the noise at the output of each receiver antenna path is additive thermal noise, and therefore that the additive noise is uncorrelated betv~n any two antenna outputs, the noise contribution to R, (n) is non-zero only on the matrix diagonal. in environments dominated by i~erference, i.e.
the i~erfa~rce power at the raxiver is much stronger than the additive receiver noise, the noise contrt~ution to Rf(n) can be neglected. The interference covariance matrix contains information about the average spatial behavior of the interference. The eigenvectors of this matrix define the average spatial directions (in MR -space) occupied by the inte~fec~m~ce. The eigcmalues of the matrix indicate the average power occupied by the interference in each the eigendirection.
3 5 The eigendirections that are associated with large eigenvalues indicate spatial directions that receive a large amount of average intwference power. The ions a~soc~d with small eigemralues indicate spatial directions that are preferable in that they receive less average interference power.
Identifying the receive covariance matrix, R, (»), is required for finding preferable RSPs. An analogous transmit covariance matrix is required for finding preferable TSPs.
Notice that we've defined R,(n) in terms of received signal samples in Equation (I). Since the received signal samples are not usually available at the transmitter, it is preferable to wo 9sro93ss pcrrt~ss~nsi6o derive the transmit covariance matrix from the receive covariance matrix. In time division duplex (TDD) syatema, the receive sad transmit covariance matrices are substantially equal when the time betovcen reception acrd danstnission is short relative to the rsxe of time variation in the channel. In frequency division duplex (FDD) systems, the transmit and 5 receive channel values are riot oon~dated with one another at any given instant in time. However, the transmit and receive covariance matrices are substantially equal is FDD
systems when su»aent tune averaging is used in the calculation of R, . There are many techniques for determining the interference covariance matrix, two of which are discussed below.
10 One int~f~eace characton approach simply averages the received arnenna signals during time periods in which the deairad transceiver is not transmitting information.
Since there is no desired signal arriving at the receiver, the interference (and noise) covariance is precisely equal to the measured sample covariance matrix, k Rl (n) = R=(n) = k~; ~ x(n~.J)z(~.l )~ .
rte, 15 In TDD systelms, one can make use of "dead-time" to collect samples from the receiver' during which time no energ~r ii om the transmitting end arrives at the receiver. Tl~ "dead-time" is approximately equal to the round trip propagation delay between the to ends of the wireless con~unications lick, and occurs when a transoaver switches from tran~ission mode to reception mode. In the above equation, k, and k~ are the burst indexes 20 corresponding to the first and last bursts received during the dead time.
Thus, the interference covariance can be ~ with no increase in overhead.
The intce covariance matrices can also be determined while the desired signal is being transmitted to the reaeiva. One approach imrolves first determining the intorferonee signal and subsequently finding the interfa~ence signal covariance. The estirmated receives 25 i~erfaen~ce is formed by subt<acting the estimated desired signal firom the total received ~5~~
w I(n,k) = s(»,k) _ H(n,k~i(n,k) .
Therdore, oncx ttx d>amel is ideatfied and the iafxHmsdion symbols deteraained, the remaining signal is considered to be interferencx. The interference covariance matrix for bin 30 n, avera~ ova K bursts is given by, k Rr(n,~E)=~ ~I(n,k)I(n,lE)x .
~.k-x+~
It is understood that when estimating the covariance taatrix, it may be desirable to filter the covariance matrix estimates. It may also be advantageous in certain embodimerns to determine an average interference covariance matrix across SOP bins. For example, within a 35 multiple access system bursts may only be received occasionally, making it difficuh to acquire a su~cient number of bursts with which to form an accurate covariance matrix for each bin.
So instead of averaging over time (a series of received bursts), a covariance matrix is formed by averaging over the SOP bins of a single burst, Rr(k)= N~I("~ku(~k)" .
It may also be preferable to estimate the inter~~ce oovariaace matrices in an alternate frequency band. This can be done using the "dead-time" approach given above.
This may be advantageous when the transceiver has the capability of choosing alternate frequency bands for communicating. Estimates of interference in alternate bands provides the foundation for an adaptive frequency hopped scheme.

WO ~ PCT/US9"l/15160 It is understood that the examples and embodiments described herein are for illustrative proposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to be included within the spirit aad purview of this application and scope of the appended claims and then full scope of equivalents. For example, much of the above discussion concerns signal processing in the context of a wireless communication system where multiple inputs or multiple outputs are accessed by multiple transmitter antenna elements or multiple receiver antenna elements. However, the present invertion is also useful in the context of wireline channels accessible via multiple inputs or multipke outputs.

Claims (22)

CLAIMS:

1. In a digital communication system including a communication channel having one or more inputs and at least one or more outputs, a method for determining characteristics of said channel based on signals received by said one or more outputs, comprising the steps of:
(a) receiving via said one or more channel outputs, at least v training symbols transmitted via a particular spatial direction of said channel, v being an extent in symbol periods of a duration of significant terms of an impulse response of a channel; and (b) applying a substantially orthogonalizing procedure to said received at least v training symbols to obtain a time domain response for said spatial direction; and (c) applying an inverse of said substantially orthogonalizing procedure to a zero-padded version of said time domain response to obtain a frequency response for said particular spatial direction.

2. The method of Claim 1 wherein said substantially orthogonalizing procedure comprises an inverse Fast Fourier Transform and said inverse of said substantially orthogonalizing procedure comprises a Fast Fourier Transform.

3. The method of Claim 1 to 2 wherein step (a) comprises receiving exactly v training symbols.

4. The method of Claim 1 further comprising the step of repeating steps (a), (b) and (c) steps for a plurality of spatial directions.

5. The method of Claim 4 wherein each of said plurality of spatial directions corresponds to transmission through one of said plurality of channel inputs exclusively.

6. The method of Claim 3 wherein said v training symbols belong to a burst of N
symbols and said characteristics are determined for said burst.

7. The method of Claim 6 further comprising the steps of repeating steps (a), (b) and (c) for successive bursts.

8. The method of Claim 7 further comprising the step of after, step (b), smoothing said time-domain response over successive bursts.

9. The method of Claim 8 wherein said smoothing step comprises Kalman filtering or Wiener filtering.

10. The method of any of Claims 1 to 9 wherein said communication channel comprises known and unknown components, wherein said effects of said known components are removed by deconvolution, and characteristics of said unknown components are determined by steps (a), (b) and (c).

11. In a digital communication system including a communication channel having one or more inputs and one or more outputs, a method for determining characteristics of said channel based on signals received via one or more channel outputs, comprising the steps of:
receiving training symbols via said channel outputs; and computing characteristics of said channel based on said received training symbols and assumptions that an impulse response of said channel is substantially time-limited and that variation of said impulse response over time is continuous.

12. In a digital communication system including a communication channel having one or more inputs and at least one or more outputs, apparatus for determining characteristics of said channel based on signals received by said one or more outputs, comprising:
a receiver system receiving via said one or more channel outputs, at least training symbols transmitted via a particular spatial direction of said channel, being an extent in symbol periods of a duration of significant terms of an impulse response of a channel;
a substantially orthogonalizing procedure processor that applies a substantially orthogonalizing procedure processor to said at least v received training symbols to obtain a time domain response for said particular spatial direction; and an inverse substantially orthogonalizing procedure processor that applies an inverse of said substantially orthogonalizing procedure to a zero-padded version of said time domain response to obtain a frequency response for said particular spatial direction.

13. The apparatus of Claim 12 wherein said substantially orthogonalizing procedure comprises an inverse Fast Fourier Transform and said inverse of said substantially orthogonalizing procedure comprises a Fast Fourier Transform.

14. The apparatus of Claim 12 wherein said receiver system receives exactly v training symbols.

15. The apparatus of Claim 12 wherein said receiver system, said substantially orthogonalizing procedure processor and said inverse substantially orthogonalizing procedure process are arranged to operate repeatedly for a plurality of spatial directions.

16. The apparatus of Claim 12 wherein each of said plurality of spatial directions corresponds to transmission through one of said plurality of channel inputs exclusively.

17. The apparatus of Claim 12 wherein said v training symbols belong to a burst of N
symbols and said characteristics are determined for said burst.

18. The apparatus of Claim 17 wherein said receiver system, said substantially orthogonalizing procedure processor and said inverse substantially orthogonalizing procedure process operate repeatedly for a plurality of bursts.

19. The apparatus of Claim 18 further comprising means for smoothing said time-domain response over successive bursts.

20. The apparatus of Claim 19 wherein said smoothing means comprises:
means for Kalman or Wiemer filtering said time-domain response over successive bursts.

21. The apparatus of any of Claims 12 to 20 wherein said communication channel comprises known and unknown components, wherein said effects of said known components are removed by deconvolution, and characteristics of said unknown components are determined.

22. In a digital communication system including a communication channel having one or more inputs and one or more outputs, apparatus for determining characteristics of said channel based on signals received via one or more channel outputs, comprising:
a receiver that receives training symbols via said channel outputs; and a processor that computes characteristics of said channel based on said received training symbols and assumptions that an impulse response of said channel is substantially time-limited and that variation of said impulse response over time is continuous.