DLMS 알고리즘의 수렴에 관한 연구

Almost-Sure Convergence of the DLMS Algorithm

In some practical applications of the LMS Algorithm the coefficient adaptation can be performed only after some fixed delay. The resulting algorithm is known as the Delayed Least Mean Square (DLMS) algorithm in the literature. There exist analyses for this algorithm, but most of them are based on the unrealistic independence assumption between successive input vectors. Inthis paper we consider the DLMS algorithm with decreasing step size .mu.(n)=n/a, a>0 and prove the almost-sure convergence ofthe weight vector W(n) to the Wiener solution W$_{opt}$ as n .rarw. .inf. under the mixing unput condition and the satisfaction of the law of large numbers. Computer simulations for decision-directed adaptive equalizer with decoding delay are performed to demonstrate the functioning of the proposed algorithm.m.