The paper studies a LMS-based adaptive algorithm with a damping force (emulating the effect of a gravitational field on a free moving object) in the weight update recursion of the LMS algorithm. The effect of the damping force to the LMS update recursion is twofold. A negative contribution results in a time-varying Leaky-type LMS algorithm behaviour with bias-free performance. A positive contribution results in improved robustness and convergence properties. The performance of the proposed algorithm is finally evaluated by means of several simulations connected to practical applications.
Reviewer: G.Di Masi (Padova)