Robust NN-based decentralized optimal tracking control for interconnected nonlinear systems via adaptive dynamic programming

• Published in: Nonlinear dynamics Vol. 110; no. 4; pp. 3429 - 3446
• Main Authors: Guo, Bin, Dian, Songyi, Zhao, Tao
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01.12.2022; Springer Nature B.V
• Subjects: Actuators; Adaptive systems; Automotive Engineering; Classical Mechanics; Control; Control systems design; Disturbances; Dynamic programming; Dynamic stability; Dynamical Systems; Energy sources; Engineering; Mechanical Engineering; Neural networks; Nonlinear control; Nonlinear systems; Optimal control; Original Paper; Proportional integral; Robust control; Tracking control; Tracking errors; Tracking systems; Vibration; Decentralized control; Optimal control; Actuator attack; Interconnected systems; Disturbances; Observer design
• ISSN: 0924-090X; 1573-269X
• DOI: 10.1007/s11071-022-07771-2

This study presents a decentralized robust optimal tracking control scheme for a class of interconnected nonlinear systems with disturbances and actuator attacks. To save the computational energy resources, an adaptive event condition in the sensor channel and an estimated state-driven event condition in the control channel are established. Based on the output triggered information, a novel decentralized proportional-integral observer (PIO) is designed aided by the neighbor information, in which the system states and lumped disturbances are estimated. By utilizing the observer information, a reliable decentralized controller is designed. In the proposed control scheme, two control parts are included. An event-based discontinuous part is designed to force the tracking error to reach the sliding mode manifold, and an optimal control part is presented to release the effects of disturbance and minimize the control cost, the estimated disturbance values are also integrated into the control scheme as a compensation part. In addition, the optimal control problem is solved based on the neural network (NN) and adaptive dynamic programming (ADP), the stability and the weight vector of the tracking system are demonstrated, and finally, the application to the interconnected nonlinear power system verifies the merit of the presented control approach.