A study of application for an adaptive lattice filter to system identification

摘要

This paper proposes a method to apply for an adaptive lattice filter to system identification for an unknown FIR system. For constructing an inverse fileter, the lattice filter is more useful than a FIR filter, because the stability of the lattice filter can be easilier discriminated than that of the FIR filter. However, a conventional method for the lattice filter is used for linear prediction analysis and it is impossible to be used for system identification. The method, proposed in this paper, can apply for lattice filter to system identification using the normalized least mean square (NLMS) algorithm. Additionally this paper shows that the lattice filter can identify the unknown FIR system when backward reflection coefficients are independent of forward reflection coefficients. Finally, this paper proposes a method that can improve of convergence speed of system identification for the unknown FIR system using an adaptive lattice filter.