Neural networks approach to adaptive FIR filtering and deconvolution problems

The authors propose neural networks to compute the weights of adaptive filters and to solve deconvolution problems. They show both analytically and by simulation that the proposed networks are guaranteed to provide results arbitrarily close to the correct values of these problems within RC time constants on the order of hundreds of nanoseconds. The proposed neural networks can provide the optimum weights under the least-square criterion during an elapsed time of only a few characteristic time constants of the circuit