Multimodel Control of Nonlinear Systems: An Integrated Design Procedure Based on Gap Metric and <i>H</i>_∞ Loop Shaping

Abstract

An integrated multimodel control design procedure with its supporting gap metric based dividing algorithm is proposed, which integrates the multimodel decomposition and the local controller design of a nonlinear system through an improved gap metric based dividing algorithm and the H∞ loop-shaping technique. In this design procedure, desired local stability, static performance, and dynamic performance are incorporated into the system decomposition and local model set determination. Hence the closed-loop performance and stability are improved. Moreover, the right local model set for multimodel controller design is obtained, which avoids local model redundancy and simplifies the multimodel controller structure. Two nonlinear chemical systems are studied to illustrate the effectiveness of the proposed procedure.