Broadband noise cancellation using a functional link ANN based nonlinear filter

This paper presents a functional link artificial neural network based nonlinear adaptive filter (FLANAF) structure suitable for extracting sharp edged signals buried under broadband noise. Here, two schemes of learning have been proposed. In the first case the delayed version of the input signal is used for training whereas in the second case the noise free input is used as the desired signal for learning. The performance of this filter in terms of computational complexity, convergence characteristics and recovery of the desired signal have been studied. Computer simulation have been carried out to compare its performance with that of the recently reported nonlinear delayed N-path adaptive FIR (NDNAFIR) filter structure. It is shown that in all compartments the proposed FLANAF structure is superior to that of NDNAFIR filter structure