Survey and tutorial on multiple model methodologies in modelling, identification and control

Abstract

Multiple model methodology is an important approach in modelling, identification and control of complicated systems with large uncertainties (parameter uncertainty or even model structure uncertainties). It agrees with the idea of 'divide and conquer' in solving engineering problems. There are two representative strategies with scheduling the multiple models, i.e., switching strategy and weighting strategy. Theses two strategies can be alternatively viewed as 'identification' of correct model/controller. Consequently, any reasonable control system design methods can be incorporated with switching or weighting methodology to formulate robust and adaptive control schemes. This survey paper gives a brief historical review of the development of multiple model adaptive estimation (MMAE) and multiple model adaptive control (MMAC), then moves focus onto the new progress of weighted MMAC (WMMAC) that has emerged in recent years including the weighting algorithm and stability analysis of WMMAC systems, based on virtual equivalent system (VES) theory.