ADAPTIVE ESTIMATION FOR PERIODIC SIGNAL ENHANCEMENT AND TRACKING

A fast algorithm for enhancement and tracking of periodic signals with fixed or slowly varying central frequency in additive noise is proposed. The algorithm is based on results from discounted least squares identification of a harmonic signal model combined with a recursive update scheme for the estimated fundamental frequency. In particular the gain sequence of the algorithm is determined by the fundamental frequency and an eligible forgetting factor. The performance of the algorithm is compared with theoretical lower bounds and it is shown that the algorithm yields good results.