Blind DFE With Maximum-Entropy Feedback
Vladimir R. Krstić, M.L. Dukić · IEEE Signal Processing Letters · 2008
This letter proposes an entropy-gradient adaptive feedback filter specially derived for the blind decision feedback equalizer with a self-optimized configuration. Using software simulations, the parametric nonlinearity of the feedback filter is aligned through two operation modes—blind acquisition and soft transition—to respond with a maximum entropy signal. As a result, a simple stochastic gradient algorithm is obtained, and it can be easily optimized for an applied signal constellation.