On the Use of Lyapunov Functions for the Design of Complex FIR Digital Filters
Yue‐Dar Jou, Fu‐Kun Chen · 2006
In this paper, a neural network based approach is extended to the design of arbitrary complex FIR filters. The minimization of difference between the frequency responses of desired FIR filter and the designed one is formulated as Lyapunov energy function. The filter coefficients can be therefore obtained by using a set of dynamic non-linear equations. As compared to linear algebra based approaches, such neural networks possess the advantages of inherent parallelism, realtime processing, high computational capability and ease of hardware implementation