Mallat algorithm of wavelet for time-varying system parametric identification

This paper presents a mallat algorithm of wavelet for identifying time-varying parameter method. Expand the time-varying parameters used mallat algorithm, then the identification of the time-varying parameters is converted to the problem of identifying invariant coefficients. Select the filter banks which make the most of the energy of time-varying parameters concentrated in low-resolution part, then select multi resolution depths, on which gives the wavelet expansions of time-varying parameters. Finally invariant coefficients are identified through least-square method, then reconstruct them and get the time-varying parameters we needed. Simulation results show the wavelet analysis method is accurate and timely tracking of mutations in the signal, showing the superiority of the algorithm.