Reliability based soft transition technique for dual-mode blind equalizers

We propose a new technique that facilitates the soft transition between a startup algorithm and a decision directed (DD) algorithm in blind adaptive equalizers. The algorithm-pair is combined using a reliability measure that is proportional to an estimate of the probability of the equalizer detecting a correct symbol. This measure takes into account both the equalized signal and its statistical distribution. The main feature of the technique is in the smooth and automatic switching from the startup algorithm to the DD algorithm and vice versa depending on the value of the reliability measure. This technique has been compared with the popular Benveniste-Goursat and stop-and-go algorithms and is shown to exhibit a faster rate of convergence and lower steady state error.