Near-optimal symbol-by-symbol detection schemes for flat Rayleigh fading

The optimal MAP detection algorithm for demodulation of M-QAM signaling in the frequency flat Rayleigh fading channel is proposed. A recursive structure is derived which allows for symbol-by-symbol metrics to be computed and data decisions to be made in an efficient manner. The complexity problem inherent to optimal detection schemes is addressed and approximations to optimal detection are considered which reduce complexity to practical levels yet still provide near-optimal performance levels. Reduced complexity algorithms are obtained through the use of decision feedback and thresholding which allow for higher order modulations such as 16-QAM to be implemented. Performance characterizations of the algorithms presented are obtained via Monte Carlo simulation and results show that near-optimal bit error probability levels can be achieved. Extensions of the algorithms to include diversity combining techniques are also developed and shown to provide improved performance. Some theoretical derivations for optimal detection of 16-QAM in fading with diversity are also presented.