과거계수 벡터를 이용한 부호 부밴드 적응 필터

초록

This paper proposes a new Sign Subband Adaptive Filter (SSAF) algorithm incorporating a reusing coefficient vector technique. The proposed method combines L1-norm minimization of the error with a squared L2-norm constraint, regulating the squared distance between current and past weight vectors to reduce the steady-state mean-square deviation compared to conventional SSAF. Simulation results demonstrate that the proposed algorithm achieves excellent convergence performance and superior tracking performance.

키워드