Adaptive blind channel identification

We present an adaptive blind channel equalizer, which is implemented by an adaptive blind channel estimator followed by a Viterbi detector. The adaptive blind channel estimator is realized by a temporal neural network which is globally convergent to the FIR channel response without any training signals. The adaptive blind channel estimator only requires either twice oversampling the received signals or signals from two antennas. Simulation shows that the adaptive blind channel equalizer without training signals performs equally well, if not better than, the decision feedback equalizer with training signals