A novel neural network-based approach for designing digital filters

A novel digital filter design approach based upon Neural Network Optimization (NNO) technique is proposed in the present paper. To demonstrate the feasibility of the NNO design approach; a Continuous Hopfield Neural Network (CHNN) model is chosen and the relation between the MSE criterion and the Lyapunov function is also established. The implementation of the NNO digital filter design is described together with some design guidelines. A few linear phase FIR filter design examples are given and the advantages of the NNO approach over conventional design techniques are illustrated.