Cascaded RLS Adaptive Filters Based on a Kronecker Product Decomposition

The multilinear system framework allows for the exploitation of identification problem from different perspectives in context various applications, such as nonlinear acoustic echo cancellation, multi-party audio conferencing, and video which could be modeled through parallel or cascaded filters. In this paper, we introduce memoryless memory structures that are described a bilinear perspective. Following structures, develop recursive least-squares algorithm by considering Kronecker product decomposition concept. We have performed set simulations aiming both long length impulse responses reverberation effect.