Output error method for identification of multiple-model process with transition

This paper is concerned with the identification of multiple model process with transition using the output error (OE) method. Local multiple linear models with output error model structure are identified at fixed operating points first and then a global nonlinear model is approximated by interpolating the multiple linear models with exponential weighting functions. With all the obtained initial values of the system's parameters, nonlinear optimization strategy is implemented to find the optimal values of the parameters for both the local multiple models and their corresponding weighting functions. Finally a more accurate global model can be obtained. The effectiveness of the proposed approach is demonstrated through simulation examples.