Model Prediction Algorithm and Co-Design of Time Delayed Networked Control Systems

Abstract

The Unmanned Aerial System (UAS) uses moderately complex, embedded hardware and software for realization of the control objectives. In such systems, there is a necessity for networking various sensors, actuators and the controllers through a wireless medium. This leads to the need for careful examination of the quality of both the network and the control in a holistic sense, since the feedback loop is closed over the wireless network. We propose a co-design of control and communication algorithms to mitigate the time delay effects and to ensure certain degree of resilience. Modified Smith predictor and the Dynamic Matrix Control are the control strategies considered. Performance error of control loop becomes larger with the increase in network time delay, which would affect system stability. IEEE 802.15.4 wireless network standard is adopted in the present work for network environment. Higher priority is assigned to time-critical task in the mission mode of aerial vehicles through the optimized tuning of wireless network protocol parameters, in proportion to the associated control loop error values. The simulation results show the effectiveness of the controller designs, and verify the co-design algorithms as well. Co-design helps to alleviate the negative effects of time delay arising in the signal communicated through the wireless network.