Adaptive algorithms for the frequency-domain identification of a second-order Volterra system with random input

Abstract

The authors describe two practical digital procedures (i.e. complex RLS and LMS adaptive algorithms) for the identification of a second-order Volterra system subject to stationary (more precisely, stationary to order four) random input through the higher-order spectral analysis of the measured I/O data. Also, the validity of the nonlinear transfer function approach is demonstrated by analyzing experimental data of the surge response of a tension leg platform (TLP) to irregular sea waves.<>