A modified hierarchical RLS algorithm for adaptive FIR filtering

In this paper, a modified hierarchical least mean square (HLMS) algorithm for adaptive FIR filtering is proposed, in which the desired responses of each subfilter in the hierarchical level, besides the bottom level, are different from one of the recently proposed HLMS algorithm. The approach proposed intends to improve the performance of HLMS algorithm by increasing the estimation precision of each subfilter. The results of simulation evaluation indicate that the modified HLMS algorithm could only maintain the feature of speeding up convergence, but also obtain lower error on the impulse response than one of the recently proposed HLMS algorithm.