Novel RLS algorithm with internal model and its stability margin

This paper deals with a new recursive least squares (RLS) algorithm for time-varying parameters using an internal model and a stability margin of the algorithm. This algorithm is given from a standpoint that the varying parameter changes can be approximated as a polynomial function of time with a finite degree. Thus, the proposed algorithm involves an internal model of the function in the adaptive loop so as to consider an influence of the varying parameters. A sufficient condition for assuring stability of the proposed algorithm is given by the output strictly passive and small-gain theorem. Furthermore, the stability margin of the algorithm is shown and a simulation is down to perform the validity of the new least mean squares algorithm.