Two-Stage Recursive Least Squares Parameter Identification for Cascade Systems with Dead Zone

Abstract

In this paper, two-stage recursive least squares algorithm (TS-RLS) is investigated for parameter identification of cascade systems with dead zone. In order to estimate the slopes and endpoints of the dead zone, switching functions are presented to reconstruct the expression of dead zone. All the parameters of linear subsystems and dead zone are separated by using the key term separation principle, which is applied to convert the cascade systems into a quasilinear model. The proposed identification algorithm not only estimates all the parameters of the cascade systems, but also reduces the computation cost of identification process. The result of the simulation illustrates the flexibility and efficiency of proposed identification algorithm.