Optimization algorithms for adaptive filtering of interferences in corrupted signal

Abstract

Neural network adaptive filters are mainly used for the interference cancellation techniques. The gradient based design methods are well developed for the design of neural network adaptive filter but they converge to local minima. This paper describes the global optimization interference cancelling techniques for adaptive filtering of interferences in the corrupted signal. The system is designed using the adaptive filtering of the interferences in the corrupted signal using the Back Propagation Neural Network (BPNN) algorithm, Genetic Algorithm (GA), and Bee Colony (BC) algorithm. These optimization algorithms are used for initialization of weights, learning parameters, activation function and selection of network structure of the artificial neural network. The adaptive filtering system is designed using an adaptive learning ability of BPNN algorithm. This paper presents a comparison of evolutionary optimization algorithm such as hybrid GA-BPNN and BC-BPNN algorithm for the interference cancellation in corrupted signals.