An adaptive inverse-QR recursive Levenberg-Marquardt algorithm

In this paper, we present the adaptive inverse-square root recursive Levenberg-Marquardt (inverse-QR RLM) algorithm. We introduce how to formulate the proposed inverse-QR RLM algorithm based on the least squares criterion for nonlinear adaptive filtering. We describe briefly about the inverse-QR RLM algorithm which can preserve the Hermitian symmetry of inverse autocorrelation matrix with the method of modified product of QR-decomposition and its Hermitian transpose. This symmetrical property leads to the benefit of parallel implementation. Simulation results show that the proposed algorithm can obtain a fast convergence rate to unknown channel estimation or system identification.