Applied library of adaptive lattice filters for nonstationary signal processing

A diversity of adaptive filters for non-stationary signal processing based on lattice structures is presented. The ability of non-stationary signal processing is achieved by means of sliding window and (or) regularization application to the estimation of the correlation matrix of processed signals. Both sequential and novel parallel implementation of the filters are considered. Cascade architecture allows the efficient implementation of the developed lattice filters in hardware or software. Efficiency of software implementation in innovative Russian “MULTICORE” DSP platform is evaluated.