Convex Combination of FXECAP– FXECLMS Algorithms for Active Noise Control

摘要

Adaptive algorithms used for active noise control systems require a high convergence speed in order to be usable in real ANC applications. Affine projection (AP) algorithms offer a high convergence rate, however, this speed entails a high misadjustment. AP algorithms have been combined with other algorithms to address this misadjustment level, these combinations are known as a convex combination. However, such combinations tend to be computationally complex which precludes them from being used in ANC applications. This work presents a new convex combination, composed of the Filtered-x Error Coded Affine Projection and the Filtered-x Error Coded Least Mean Square Algorithm (CC FXECAP-FXECLMS), with the aim of obtaining an algorithm that retains the speed of the AP while achieving misadjustment levels of the LMS, while at the same time producing an algorithm with a computational complexity low enough to be practical for use in real ANC applications. Test results have confirmed the algorithm's capabilities and proven that the algorithm is suitable for this role.