The inverse QRD set-membership RLS algorithm

This paper introduces a robust QR-decomposition version of the Set Membership RLS (SM-RLS) algorithm. The SM-RLS algorithm was recently proposed to improve the numerical stability of the so-called BEACON algorithm, while keeping the same performance. Our algorithm takes the numerical stability a step further, by considering the Cholesky factor of the input autocorrelation matrix and using it in the Givens rotation based update equations. The performance of the proposed algorithm in comparison to the SM-RLS algorithm is verified through computer simulation. The results show a superior performance of the proposed algorithm as compared to the original SM-RLS in finite precision environment, without any compromise on the convergence and tracking performance.