Blind correlated TDL equalization in fast Rician fading channels

A new blind technique is presented which combines channel estimation and data recovery for transmission over the rapidly changing, Rician frequency-selective channel with unknown fading statistics. At each symbol interval, a decision algorithm first makes a primary data estimate based on the constrained linear MMSE criterion. Then the new primary data estimate along with the few previous ones is used by subsequent MMSE channel estimation. Finally a linear equalizer is adapted to recover the data on a symbol-by-symbol basis, which is therefore suitable for wireless relay systems. The error performance is evaluated by simulation using the estimated signal MMSE and BER as metrics, and is very promising.