Nonlinear signal reconstruction based on recursive Moving Window Kernel Method

Reconstruction problem for signals generated by discrete nonlinear dynamic system is considered via unified approach to recurrent kernel-based dynamic systems. In order to prevent the model complexity increasing under on-line identification, the reduced order model kernel method is proposed and proper recurrent Least-Square identification algorithms are designed along with conventional regularization technique. The recurrent version of Moving Window Kernel Method is also considered and suitable identification algorithm is developed, which has tracking properties and may be successfully used for on-line identification of nonlinear and nonstationary signal reconstruction.