Comparative study of LMS and L∞-norm adaptive turbo equalization

In this paper we show that adaptive turbo receiver system employing adaptive least mean square (LMS) cost function based decision feedback equalizer (DFE) has a better bit error rate (BER) performance compared to L_∞-norm cost function based adaptive turbo DFE. The turbo system employs an adaptive turbo DFE concatenated with a log-maximum a-posterion (MAP) decoder, recursively exchanging statistical information between the decoder and the equalizer. A quasi-stationary, time-invariant mobile inter symbol interference communication channel is considered. Monte Carlo simulations have shown that with large number of filter-taps in Turbo-DFE, there is a reduction in BER performance for adaptive LMS based systems compared to _∞-norm based turbo system.