Fuzzy identification of nonuniformly sampled nonlinear systems based on forwards recursive input–output clustering

Based on forwards recursive input–output data clustering, a recursive least squares (RLS) algorithm is proposed to estimate nonuniformly sampled nonlinear systems. The relationship of linear and nonlinear systems is studied under nonuniform sampling, and a fuzzy model is constructed for the global system. Forwards recursive input–output data clustering based on k-means clustering is used to identify the fuzzy rule number and the antecedent parameters. Based on the membership function, the consequent parameters are identified by an RLS algorithm. A practical application verified the efficiency of the method.