Weighted Filtered-s LMS Algorithm for Nonlinear Active Noise Control
Nonlinearity in Active Noise Control (ANC) must be taken into account for the improvement in performance of the system. Artificial neural network topologies are efficient techniques to deal with nonlinear modelling. Emergence of Functional Neural Network (FLANN) heralded a new paradigm in nonlinear ANC. Employment of FLANN in nonlinear ANC is characterized by simple structure, relatively low in computational load and slow convergence capability. In this paper, the structure of FLANN is modified by introducing a set of combining weights. The objective is to accelerate the slow convergence of ANC. The algorithm to update the FLANN and newly introduced combining weights are also proposed. The proposed structure and algorithm have faster convergence capability without losing the FLANN’s structural simplicity. This modified structure achieved its objective without compromising in Mean square error performance and a substantial increase in computational requirement. Exhaustive computer simulation experiments are also presented to validate the efficacy of the proposed algorithm.
KeywordsNonlinear active noise control Filtered-s LMS algorithm Weighted Filtered-s LMS algorithm Functional link artificial neural network
- 1.Kuo SM, Morgan DR (1996) Active noise control systems: algorithms and DSP implementations. Wiley, New YorkGoogle Scholar
- 2.Wan EA (1996) Adjoint LMS: an efficient alternative to the Filtered-x LMS and multiple error LMS algorithm. In: Proceedings of IEEE ICASSP, pp 1842–1845Google Scholar
- 5.Das DP, Panda G (2004) Active mitigation of nonlinear noise process using novel Filtered-s LMS algorithm. IEEE Trans Speech Audio Process 12(3):313–322Google Scholar
- 14.Xhao H, Zeng X, He Z, Li T (2012) Adaptive extended pipelined second order Volterra filter for nonlinear active noise controller. IEEE Trans Audio Speech Lang Process 20(4):1394–1399Google Scholar
- 15.Zhao H, Zeng X, He Z, Yu S, Chen B (2016) Improved functional link artificial neural network via convex combination for nonlinear active noise control. Appl Acoust 42:351–359Google Scholar