Performance Evaluation of Adaptive Filters Using a Novel Technique of Matrix Inversion

Adaptive filtering is one of the emerging areas in the field of digital signal processing. Adaptive filter alters its coefficients according to the situations, in order to minimize the cost function. The cost function is the difference between a desired or reference signal and the filter output signal. Once the desired signal is available for the process, the cost function is minimized by the adaptive filter and the tap weights are calculated algorithmically. According to the wiener solution the filter tap weight vector is calculated by multiplying the inverse of auto-correlation matrix R with the cross-correlation vector p. This paper evaluates the performance of adaptive filters by taking the inverse of the correlation matrix R through a new algorithm. The algorithm is tested and implemented in adaptive filtering for the applications of system identification and noise cancellation. Various parameters of adaptive filtering are examined for the processes. All the simulations are done in MATLAB software.