Statistical and neuro-computing channel equalization approaches for MIMO

The performance comparison of two different channel estimation techniques for Rayleigh flat fading time varying Multiple Input Multiple Output (MIMO) channels are reported in this work. The first technique discussed here is the Autoregressive (AR) model, a statistical method of channel estimation and the second scheme is the channel estimation based on a dynamic class of Artificial Neural Network (ANN), the Focussed Time Delay Neural Network (FTDNN). Zero-forcing (ZF), Minimum Mean Squared Error (MMSE) and Maximum Likehood (ML) channel equalization techniques are employed for both the estimation techniques separately to investigate the performance of the system. Simulations are performed for different International Telecommunication Union (ITU) specified channel conditions and certain conclusions are drawn from them.