Combinational neural net for adaptive filtering

An electronic circuit which is based on the idea of combinational neural nets to perform adaptive filtering is presented. This filter can get the optimum filter coefficients in each sampling period since the neural net has fast parallel calculation ability. Therefore, the adaptive speed of the filter is very fast. It is shown by analysis that the network is guaranteed to obtain the correct coefficients of the optimum filter within time tau =RC. The simulation results of the filter by SPICE are shown.<>