Abstract

Multiple-Input Multiple-Output (MIMO) systems that provide significant increase in channel capacity is rapidly emerging as the new frontier of wireless industry. MIMO systems require the simultaneous use of multiple transmit and receive antennas to dramatically increase data rates and to improve performance reliability. An effective and practical way to approach the capacity promised by MIMO systems is to employ space-time coding (STC). It elegantly combines temporal and spatial correlation into the transmitted symbols to realize diversity and coding gains. Most STC schemes are designed for known quasi-static channels however this assumption is not always justified. MIMO channels often undergo frequency selective fading that leads to intersymbol interference (ISI), which limits the performance of MIMO systems. The effect of imperfect channel estimation on the bit error rate (BER) of MIMO systems utilizing STC is investigated. An analysis and comparison into the BER degradations of simple transmit diversity (STD) and maximal ratio combining (MRC) schemes due to multipath channel estimation errors are presented. Closed form expressions are derived for the BER performances of the schemes that employ an equalization process to mitigate the ISI caused by the multipath in frequency selective channel. BER curves show that the performance deterioration in the MIMO scheme outweighs the benefits achieved over the single antenna case when the channel estimation errors are large. Results expose the deleterious effects of inaccurate channel estimation on the performance of MIMO systems. Hence, the development of practical and novel channel estimation approaches are desired for MIMO systems using STC. This dissertation introduces a new MAP based channel estimation technique that is amenable to STC scheme employing two transmit antennas and operating in multipath bandlimited channel. The complex channel parameters are treated as two real-valued tap coefficients; each taking one of M possible amplitude levels with equal probability. The proposed estimation technique is based on an iterative procedure derived through the maximum a posteriori (MAP) probability approach. Unlike classic estimation techniques, we iterate on the probabilities of the different coefficients rather than on the values of the coefficients. Two low complexity algorithms based on the developed channel estimation technique and simple to implement in practical MIMO systems are also introduced. The performances of the two algorithms are assessed by combined analysis and simulation. Results are presented and compared against the performance of conventional channel estimation techniques. Results show that the required performance can be achieved with less number of iterations using the proposed algorithms compared to conventional techniques.