Nonlinear Predictive Control with a Gaussian Process Model：用高斯过程模型的非线性预测控制精品.pdfVIP

Nonlinear Predictive Control with a Gaussian Process Model Juˇs Kocijan1,2 and Roderick Murray-Smith3,4 1 Jozef Stefan Institute, Jamova 39, SI-1000 Ljubljana, Slovenia jus.kocijan@ijs.si 2 Nova Gorica Polytechnic, Nova Gorica 3 Dept. Computing Science, University of Glasgow, Glasgow 4 Hamilton Institute, National University of Ireland, Maynooth Abstract. Gaussian process models provide a probabilistic non-para- metric modelling approach for black-box identification of nonlinear dy- namic systems. The Gaussian processes can highlight areas of the in- put space where prediction quality is poor, due to the lack of data or its complexity, by indicating the higher variance around the predicted mean. Gaussian process models contain noticeably less coefficients to be optimized. This chapter illustrates possible application of Gaussian process models within model predictive control. The extra information provided by the Gaussian process model is used in predictive control, where optimization of the control signal takes the variance information into account. The predictive control principle is demonstrated via the control of a pH process benchmark. 1 Introduction Model Predictive Control (MPC) is a common name for computer control algo- rithms that use an explicit process model to predict the future plant response. According to this prediction in the chosen period, also known as the prediction horizon, the MPC algorithm optimizes the manipulated variable to obtain an optimal future plant response. The input of chosen length, also known as control horizon, is sent to