Experimental Study on Enhanced FXLMS Algorithm for Active Impulsive Noise Control

Active noise control (ANC) technique with the filtered-x least mean square (FXLMS) algorithm has proven its efficiency and drawn increasingly interests in vehicle noise control applications. However, many vehicle interior and/or exterior noises are exhibiting non-Gaussian type with impulsive characteristic, such as diesel knocking noise, injector ticking, impulsive crank-train noise, gear rattle, and road bumps, etc. Therefore, the conventional FXLMS algorithm that is based on the assumption of deterministic and/or Gaussian signal may not be appropriate for tackling this type of impulsive noise. In this paper, an ANC system configured with modified FXLMS (MFXLMS) algorithm by adding thresholds on reference and error signal paths is proposed for impulsive noise control. To demonstrate the effectiveness of the proposed algorithm, an experimental study is conducted in the laboratory. Various impulses with different durations are added in the Gaussian noise to simulate the typical impulsive noise. A series of low pass filters (LPFs) are used to model the various transmission paths, where the multiple reference signals are generated and used for the MFXLMS algorithm. Experimental results demonstrate that the enhanced MFXLMS algorithm exhibits improved robustness and performance for impulsive-type noise control as compared to the conventional FXLMS algorithm, and promising reductions at various impact events are achieved without causing instability.