Adaptive equalization techniques for the mitigation of multipath effects in OFDM based wireless transmission systems

Broadband wireless access is rapidly gaining ground as the technology of choice for voice and high speed data communications. The wireless channel is distinct and unpredictable because of factors such as shadow fading, multipath fading, Doppler spread and time dispersion. Due to the time-varying nature of radio channels, their multipath characteristics, and the demand for high transmission data rates, equalization schemes with fast convergence times are an essential requirement for conservation of end-to-end bit error rate. Several techniques exist to reciprocate the effect of these channel impairments. One such technique for adaptive equalization using quantized state methods, developed by Kumar and Moore, has been studied and its characteristics have been compared and contrasted with conventional equalization methods. The compromise between counter-active equalizer attributes of implementation complexity and convergence rate has been analyzed through simulation. A reference system based on an 802.11a OFDM modem has been used as a vehicle to demonstrate the effectiveness of these novel equalization methods in the presence of artifacts peculiar to a 5GHz mobile radio channel. Despite reduced convergence rates these equalization techniques provide extremely competitive bit error rate performance at a fraction of the original implementation cost. Simulation results also elucidate significant bandwidth conservation by allowing the system to operate at effectively double the specification prescribed data rates with only slight bit error deterioration. The gargantuan computational effort required by least-squares based recursive adaptive algorithms makes them extremely prohibitive for commercial applications. In this context the quantized state techniques present themselves as an ideal candidate for channel equalization in wireless systems.