Filtered-X Radial Basis Function Neural Networks for Active Noise Control

Journal of ICT Research and Applications

Budi Sulistyo¹  Dimitri Mahayana¹  Carmadi Machbub¹  Budi Rahardjo¹ 

关键词: Active Noise Control;    Adaptive Nonlinear Control;    DSP;    Radial Basis Function Networks.;   

DOI  :  10.5614/itbj.eng.sci.2004.36.1.2

学科分类：电子、光学、磁材料

来源: Institute for Research and Community Services ITB

【 摘 要 】

This paper presents active control of acoustic noise using radial basis function (RBF) networks and its digital signal processor (DSP) real-time implementation. The neural control system consists of two stages: first, identification (modeling) of secondary path of the active noise control using RBF networks and its learning algorithm, and secondly neural control of primary path based on neural model obtained in the first stage. A tapped delay line is introduced in front of controller neural, and another tapped delay line is inserted between controller neural networks and model neural networks. A new algorithm referred to as Filtered X-RBF is proposed to account for secondary path effects of the control system arising in active noise control. The resulting algorithm turns out to be the filtered-X version of the standard RBF learning algorithm. We address centralized and decentralized controller configurations and their DSP implementation is carried out. Effectiveness of the neural controller is demonstrated by applying the algorithm to active noise control within a 3 dimension enclosure to generate quiet zones around error microphones. Results of the real-time experiments show that 10-23 dB noise attenuation is produced with moderate transient response.

【 授权许可】

Unknown