Abstract—In this paper we discuss in detail a recentlyproposed kernel-based version of the recursive least-squares(RLS) algorithm for fast adaptive nonlinear filtering. Unlikeother previous approaches, the studied method combinesa sliding-window approach (to fix the dimensions of thekernel matrix) with conventional ridge regression (to improvegeneralization). The resulting kernel RLS algorithm isapplied to several nonlinear system identification problems.Experiments show that the proposed algorithm is able tooperate in a time-varying environment and to adjust toabrupt changes in either the linear filter or the nonlinearity.

Index Terms—kernel methods, kernel RLS, sliding-window,system identification

Cite: Steven Van Vaerenbergh, Javier V´ıa， and Ignacio Santamar´ıa, "Nonlinear System Identification using a New Sliding-Window Kernel RLS Algorithm," Journal of Communications, vol. 2, no. 3, pp. 1-8, 2007.