Separate-variable adaptive combination of LMS adaptive filters for plant identification
ABSTRACT The Least Mean Square (LMS) algorithm has become a very popular algorithm for adaptive filtering due to its robustness and simplicity. An adaptive convex combination of one fast a one slow LMS filters has been previously proposed for plant identification, as a way to break the speed vs precision compromise inherent to LMS filters. In this paper, an improved version of this combination method is presented. Instead of using a global mixing parameter, the new algorithm uses a different combination parameter for each weight of the adaptive filter, what gives some advantage when identifying varying plants where some of the coefficients remain unaltered, or when the input process is colored. Some simulation examples show the validity of this approach when compared with the one-parameter combination scheme and with a different multi-step approach.
[show abstract] [hide abstract]
ABSTRACT: Combination approaches provide an interesting way to improve adaptive filter performance. In this paper, we study the mean-square performance of a convex combination of two transversal filters. The individual filters are independently adapted using their own error signals, while the combination is adapted by means of a stochastic gradient algorithm in order to minimize the error of the overall structure. General expressions are derived that show that the method is universal with respect to the component filters, i.e., in steady-state, it performs at least as well as the best component filter. Furthermore, when the correlation between the a priori errors of the components is low enough, their combination is able to outperform both of them. Using energy conservation relations, we specialize the results to a combination of least mean-square filters operating both in stationary and in nonstationary scenarios. We also show how the universality of the scheme can be exploited to design filters with improved tracking performance.IEEE Transactions on Signal Processing 04/2006; · 2.63 Impact Factor