Accurate error rate estimate using moment method for optimum diversity combining and MMSE equalisation in digital cellular mobile radio

Research is presented which provides fast and accurate performance analysis for optimum diversity combining and equalisation with quadrature amplitude modulation (QAM) over mobile radio channels with frequency-selective fading. The performance of optimum diversity combining and equalisation over a noise-limited channel has been presented previously, but the error rate calculation was based on the upper bound method. The accurate performance of such a diversity receiver has not been evaluated completely. An analysis is presented in which the upper bound method for each individual run is replaced by the moment method in a quasi-analytical (QA) Monte Carlo simulation. Simulation techniques are first used to evaluate the residual ISI after equalisation, and then moments of this ISI are calculated using a very fast and accurate recursive method of computation. Based on these moments, the author uses the series expansion to obtain the error rate of the system. The multipath channel is treated as quasistatic and the system, including the channel, is assumed to be linear. Minimum mean-square-error (MMSE) equalisation techniques with linear and decision-feedback structures are considered and compared in the analysis. A method is defined for finding the transfer function of the infinitely long decision feedback equaliser, which requires greater computational complexity. Numerical results show that by using the moment method the performance can be improved significantly in comparison to the upper bound approach