Robust Fixed-Time Sliding Mode Attitude Control for a 2-DOF Helicopter Subject to Input Saturation and Prescribed Performance

Hao Shen,Xin Yu,Huaicheng Yan,Ju H. Park,Jing Wang
DOI: https://doi.org/10.1109/tte.2024.3402316
IF: 6.519
2024-01-01
IEEE Transactions on Transportation Electrification
Abstract:This paper explores the issue of robust attitude control for a 2-DOF helicopter system under the fixed-time control rule. Thanks to the reinforcement-learning strategy, the optimization results for the attitude control objective have been achieved. Under the basic framework of the Actor-Critic Neural Networks, this paper not only solves a better solution of the cost-to-go function but also successfully estimates the external disturbance torque existed in the 2-DOF helicopter system. Furthermore, in conjunction with a sliding mode switching mechanism and a novel reaching law, this study introduces a new approach for effectively accomplishing the objective of attitude control while adhering to the constraints of input saturation and prescribed performance. Compared with other types of controllers, a fact can be validated that it has a better action performance of attitude control. In particular, under the action of the controller, each state variable has a stable bound over a specific fixed time. Finally, simulation and comparison examples offer evidence to demonstrate that the proposed control technique is stable.
What problem does this paper attempt to address?