Algorithm Of Active Noise Control For Combating Impulsive Noise

Nowadays, noise pollution is as known one of the four major pollutions. To solve the problem that the increasingly serious problem of noise pollution and the drawback of the conventional passive noise control method at low frequency, active noise control methods have received increasingly large amounts of attention. At present, active noise control theory has been studied extensively, but it still cannot reach the practical engineer application standard because several practical problems do not be considered. For example, when noise is corrupted by impulsive noise, it is difficult to achieve an ideal noise reduction performance using conventional active noise control method, or it may fail to work. However, in many practical application areas, such as the ICU ward ventilator, infant incubators, pumps, and crushers have impulsive characteristics. To obtain good robustness against impulsive noise and overcome the limitation of the second-moments-based algorithm, this paper proposed several adaptive active impulsive noise control algorithms, which have fast convergence rate, small steady-state error, and low computational complexity. The main contributions of this paper is outlined as follows:Firstly, this paper briefly introduce the development of active noise control theory, state of research, and three main structures (feedforward active noise control system, feedback active noise control system, and multi-channel active noise control system). Then, we analyzed the advantages and drawbacks of such structures and their working environments. In addition, the mathematical models of impulsive noise are analyzed in detail.Secondly, this paper review several commonly used active noise control algorithms, namely, the filtered-x least mean square (FxLMS) algorithm, filtered-x least mean p-novm (FxLMP) algorithm, filtered-x logarithm-based least mean square (FxlogLMS) algorithm, and filtered-x M-Estimate algorithm (FxLMM). Also, we analyzed the superiority and disadvantage of these algorithms and their performances in the presence of impulsive noise.Thirdly, based on the FxlogLMS algorithm, an improved version (CRFxLMS) algorithm is proposed via the convex combination approach. As compared with the FxlogLMS algorithm, this algorithm has an improved performance in terms of convergence rate and steady-state error. Simulations and experiments verified the performance of the algorithm.Finally, this study developed an improved active noise control algorithm, called the filtered-x least mean square-based convex combination maximum correlation entropy algorithm(CMCCFxLMS), which is derived from the maximum correlation entropy (MCC). In the previous studies, LMS algorithm based on the MCC has good robustness combating impulsive noise. To further improve its performance, we introduced the convex combination approach and developed an improved method for adaptation of the mixing parameter. Therefore, the proposed algorithm holds the fast convergence speed of large step-size filter and low steady-state error of small step-size filter. Simulation results are presented to evaluate the performance of the proposed algorithm.