IDENTIFICATION OF NON-UNIFORMLY SAMPLED MULTIRATE SYSTEMS USING ORTHONORMAL BASIS FILTERS

In this work, we propose a novel scheme for identifying fast-rate deterministic and disturbance model from irregularly sampled multirate input-output data. Generalized Orthogonal Basis Functions (GOBF) are used for representing model dynamics. In the first step, we identify MISO output error (OE) type models using irregularly sampled outputs and inputs changing at the fast rate. To capture the dynamics of unmeasured disturbances, we propose to develop AR type models with time varying coefficients using the irregularly sampled model residuals generated in the first step. The efficacy of the proposed identification technique is demonstrated by simulation studies on Shell control problem.