Plenary Speech 5

Plenary  Speech 5 : Fast Model Predictive Control for Nonlinear Systems and Nonlinear Predictive PID

Dr Wen-Hua Chen (IEEE Fellow) holds Professor in Autonomous Vehicles and the founding Director of Centre for Autonomous Systems in the Department of Aeronautical and Automotive Engineering at Loughborough University, UK. Prof. Chen has a considerable experience in control, signal processing and artificial intelligence and their applications in aerospace, automotive and agriculture systems. In the last 20 years, he has been working on the development and application of unmanned aircraft system and intelligent vehicle technologies, spanning autopilots, situational awareness, decision making, verification, remote sensing for precision agriculture and environment monitoring. He is a Chartered Engineer, and a Fellow of IEEE, the Institution of Mechanical Engineers and the Institution of Engineering and Technology, UK.  Prof Chen currently holds a 5-years Established Career Fellowship of the UK Engineering and Physical Sciences Research Council (EPSRC) in developing AI enabled control systems for robotics and autonomous systems.

Abstract: Model Predictive Control (MPC) provides a promising mechanism to realise numerical optimal solutions online to achieve best possible performance. It is becoming the most widely used advanced control technology. With its penetration from the traditional chemical process into systems with fast dynamics such as power electronics and motor drivers, real time implementation attracts significant attention since online optimisation has to be conducted in MPC. This talk introduces fast MPC for nonlinear systems with fast dynamics. It shows that, by using Taylor series expansion, it is possible to develop an analytic solution of nonlinear MPC where online optimisation is not required, which can be implemented on any systems with fast dynamics easily. This group of nonlinear MPC algorithms has a number of promising properties. First, its stability could be easily established.  More interestingly, its performance such as overshoot and transient period could be directly specified. By combining the analytic nonlinear MPC with a nonlinear disturbance observer, we show that a nonlinear predictive PID could be derived where, in addition to conventional nonlinear PID terms, there is an extra term based on prediction of system behaviour described by the nonlinear model.