A comparison of leaky RLS and its approximation algorithms for an system

摘要

本報告では能動騒音制御システムへの応用のためのfiltered-X型のleaky RLSとその近似アルゴリズムを述べ、それらの収束特性を比較する。両者のアルゴリズムには収束特性と演算量の間にトレードオフの関係があり、シミュレーション結果によりLRLSアルゴリズムがその近似アルゴリズムより収束速度が高いことが示される。%This report presents a comparison between the filtered-X leaky recursive least squares (FXLRLS) algorithm and filtered-X affine projection (FXAP) algorithm and their properties have been exploited for understanding the filters performance in an active noise control (ANC) systems An analysis of these two algorithms can produce good tradeoff between the computational complexity and convergence speed. We discuss the steady-state and transient behavior of conventional and modified version of AP algorithm and it has been shown from the simulation results that convergence speed of LRLS is lightly faster than the AP algorithm.