Performance analysis of MIMO-OFDM systems

The design and development of reliable, high data rate, wireless links is a challenging proposition from a system design perspective. There is significant emphasis on research related to spectrally efficient communication systems as well as multiple antenna systems, in an effort to achieve a multifold increase in throughput. MIMO-OFDM (Multiple Input Multiple Output, Orthogonal Frequency Division Multiplexing) is a promising new technology that combines the spectral efficiency of OFDM and the throughput gains of MIMO systems. Theory and simulations show remarkable gains in performance with this technology. Essential to proving any new technology is the development of testbeds that provide a realistic environment for validating gains reported by theory and simulations.; Analog impairments significantly degrade the performance of any practical wireless communication system. Two of the most common analog impairments are I/Q mismatch and phase noise. The impact of I/Q mismatch and phase noise in the context of single carrier systems and single antenna OFDM systems has been studied extensively in the literature. However, their impact on MIMO-OFDM systems has hitherto not been explored. This dissertation analyzes the impact of I/Q mismatch and phase noise on the performance of MIMO-OFDM systems as the number of antennas is increased. An RLS based adaptive algorithm for joint MIMO decoding and I/Q mismatch cancellation is derived and the performance of the MIMO-OFDM system with I/Q mismatch cancellation is analyzed.; The multi-antenna testbed developed at UCLA is described along with all the communication algorithms. The calibration of the testbed in an AWGN channel is discussed and its performance is compared with theory and simulations. Indoor and outdoor field measurements conducted with this testbed are analyzed and compared with simulations. The performance of the MIMO-OFDM system in line of sight (LOS) and non line of sight (NLOS) channels and in outdoor channels is evaluated. The impact of antenna spacing on the performance of the system is explored as well. Finally, the optimum packet overhead in terms of the training symbols for channel inverse estimation and continuous pilots for carrier frequency offset estimation and tracking are explored.