Multichannel Filtered-X Set-Membership Affine Projection-Like Algorithm

The processing of several channels using active noise control (ACN) systems demands high performance architectures due to the computational complexity involved. In this paper, we have proposed a filtered-x set membership afinne projection-like algorithm (FXSMAP-L) along with an evolutionary methodology to improve the computational cost by keeping a high convergence speed. In particular, our efforts were dedicated to reducing the number of multiplications involved in the proposed algorithm since multiplication is one of the most demanding operation in terms of processing speed and area-power consumption. The results presented demonstrate that the number of updates were reduced, compared to other reported solutions based on filtered-x set membership schemes. This potentially will allow the implementation of the proposed algorithm in advanced hardware systems used in health care, automotives, etc., with efficient power, area and performance utilisation in processing multiple channels.