Self-Configurable Neural Network Processor for FIR Filter Applications

Abstract

A self-configurable system is one that is designed primarily for the purpose of reconfigurable control and adaptive signal processing. It evolves by restructures and readjustments back and forth which can track the environment and the system variation in time. Processing methods and application areas include but not limited to transmission enhancement such as filtering, equalization, and noise cancellation. The performance of our proposed self-configurable neural network processor (SCNNP) for finite impulse response (FIR) filter are compared with those of the classical FIR filters and the traditional adaptive FIR filters. The SCNNP is an autonomous system which does not need human design knowledge of the FIR filter