In many application of noise cancellation, the changes in signalcharacteristics could be quite fast. This requires the utilization of adaptivealgorithms, which converge rapidly. Least Mean Squares (LMS) adaptive filtershave been used in a wide range of signal processing application. The RecursiveLeast Squares (RLS) algorithm has established itself as the "ultimate" adaptivefiltering algorithm in the sense that it is the adaptive filter exhibiting thebest convergence behavior. Unfortunately, practical implementations of thealgorithm are often associated with high computational complexity and/or poornumerical properties. Recently adaptive filtering was presented that was basedon Matching Pursuits, have a nice tradeoff between complexity and theconvergence speed. This paper describes a new approach for noise cancellationin speech enhancement using the new adaptive filtering algorithm named fastaffine projection algorithm (FAPA). The simulation results demonstrate the goodperformance of the FAPA in attenuating the noise.
展开▼
