Performance evaluation of detection algorithms for MOMI OFDM systems

Includes abstract.Includes bibliographical references (leaves 79-86).Introduction of Multi Input Multi Output (MIMO) Orthogonal Frequency Division Multiplexing (OFDM) as the base air interface method for Next Generation Network (NGN) will face a number of challenges from hostile channel conditions to interference from other users. This would result in an increase of detection complexity required for mobile systems. Complex detection will reduce the battery life of mobile devices because of the many calculations that have to be done to decode the signal. Very powerful detection algorithms exist but they introduce high detection complexity. NGN will employ different MIMO systems, but this research will consider spatially multiplexed MIMO which is used to improve the data rate and network capacity. In NGN different multi access modulation schemes will be used for uplink and downlink but they both have OFDM as the basic building block. In this work performance of MIMO OFDM is investigated in different channels models and detection algorithms. A low complexity detection scheme is proposed in this research to improve performance of MIMO OFDM. The proposed detection scheme is investigated for different channel characteristics. Realistic channels conditions are introduced to evaluate the performance of the proposed detection scheme. We analyze weaknesses of existing linear detectors and the enhancements that can be done to improve their performance in different channel conditions. Performance of the detectors is evaluated by comparison of Bit Error Rate (BER) and Symbol Error Rate (SER) against signal to noise ratio (SNR). This thesis proposes a detector which shows a higher complexity than linear detectors but less than Maximum Likelihood Detector (MLD). The proposed detector shows significant BER improvement in all channel conditions. For better performance evaluation this work also investigates performance of MIMO OFDM detectors in realistic channels like Kronecker and Weichselberger channel models