Fractional tap-length algorithm with uniformly adjusted parameters

The error width and the tap-length adaptation step-size are two critical factors that affect the overall performance of the fractional variable tap-length algorithm. Some reasonably adjusted strategies can improve the performance. However, the existing strategies for adjusting these two parameters have many problems, such as overestimation, large fluctuations, underestimation, insufficient analysis of the parameter and high complexity. Therefore, this paper proposes an efficient and unified adjustment method that utilizes the accumulated and transient full-partial errors to adjust these two parameters simultaneously. The detailed parameter ranges and specific performance of the proposed algorithm were all analyzed. In the simulation, high and low noise situations, sudden environmental changes, large- and small-scale systems, continuously changing environment and key performance curves were tested. The results of theoretical analysis and experimental tests jointly confirmed that the proposed algorithm can quickly and reasonably update the tap-length and improve some performance.