This thesis deals with performance limits and transmissiontechniques for a wireless communication link where at least thetransmitter is equipped with an antenna array and moreover hasaccess to possibly imperfect channel state information.
An antenna array on the transmit side provides the systemwith an extra spatial dimension that can be utilized for codingboth in the spatial as well as the temporal domain. The recentdevelopment of such space-time codes shows that there are waysof exploiting multiple transmit antennas while completelyavoiding traditional beamforming techniques' need of accuratechannel state information. In practice however, the transmitterusually has access to some information about the current stateof the channel. The available channel side information can thenbe used to improve the performance beyond what is possibleusing only conventional space-time codes. This, together withthe need for reliable and fast communication, providesmotivation for the work herein which shows how previousspace-time coding concepts can be extended to take advantage ofeven non-perfect channel knowledge at the transmitter.
Performance limits are investigated using tools frominformation theory. An expression for the channel capacity forthe wireless link under consideration is presented. Oneimportant result is that adjusting the output of a conventionalspace-time encoder by means of a transmit weighting matrix thatonly depends on the channel side information constitutes acapacity achieving transmitter structure. Computationalprocedures for evaluating the capacity expression areconsidered and used to obtain numerical results illustratingthe gains due to channel knowledge.
The other parts of the thesis are devoted to devisingpractical methods for exploiting channel knowledge inconjunction with space-time coding. A new performance criterionis developed that takes the quality of the channel sideinformation into account. Motivated by the optimality ofseparate space-time coding and transmit weighting, theperformance criterion is used for determining a suitabletransmit weighting matrix that adapts a predeterminedorthogonal space-time block code (OSTBC) to the availablechannel side information. The result is a low-complexityweighted OSTBC transmission scheme providing a seamlesscombination of the normally complementary strengths offered byconventional beamforming and OSTBC.
Scenarios in which the channel side information takes theform of quantized channel estimates obtained from a feedbacklink are also considered. The channel feedback is assumed tosuffer from quantization errors, feedback delay and bit-errorsintroduced by a noisy feedback channel. Methods to design thequantizer in the feedback link so as to mitigate all theseerrors are investigated. By introducing heuristic modificationsof our previously developed transmission technique, it is shownhow robustness against all three types of channel feedbackimpairments may be achieved.
To avoid the use of heuristics in case of quantized channelside information, yet another new performance criterion isdeveloped specifically for the problem at hand. Based on theperformance criterion, a procedure for utilizing the availableside information in the design of unstructured space-time blockcodes is proposed. These codes offer maximal design freedom atthe expense of an increased decoding complexity. Properties ofthe resulting codes are investigated both analytically andexperimentally. The codes outperform corresponding OSTBCschemes even when no channel knowledge is available at thetransmitter.
In addition to unstructured codes, closely relatedtechniques based on the same performance criterion are used fordesigning some linear dispersive space-time block codes as wellas designing suitable transmit weighting matrices for weightedOSTBC. An interesting observation that deserves further studyis that the design procedure for linear dispersive codes incase of no channel knowledge at the transmitter appears toautomatically produce orthogonal space-time block codes, if theparameters under consideration allow it.
Stockholm: Signaler, sensorer och system , 2003. , x, 259 p.
Beamforming, Channel feedback, MIMO, Space-time coding, Wireless Communication