Abstract: For the cancellation of impulsive noise in the desired signal, a modulated LMP algorithm is applied. According to the estimate of input signals, this algorithm is modified on the basis of the LMP algorithm. A new measure for signal distortion in impulsive noise, namely fractional order signal-to-noise ration(FSNR), is introduced to quantify the performance of the impulsive noise cancellation algorithm. The LMP algorithm and the modified algorithm are compared in performance in seven different input SNRs. In comparison with LMP algorithm, the MLMP algorithm is more stable and its convergence is faster. Results show that the performance of the modified algorithm is obviously superior to that of the LMP algorithm.

Key words: adaptive noise canceller, SNR, impulsive noise, α-stable distribution