Adaptive neural smooth finite-time controller for large-scale stochastic nonlinear state-constrained systems with feasibility removal and time delays

作者：Si, Wenjie^*; Dong, Xunde; Dong, Yanfei

来源：INTERNATIONAL JOURNAL OF SYSTEMS SCIENCE, 2022, 53(9): 1830-1847.

DOI：10.1080/00207721.2022.2025501

• Backstepping control
• stochastic systems
• neural networks
• smooth fast finite-time control
• full-state constraints
• time-delay interconnection

This paper studies the adaptive smooth finite-time neural controller design for large-scale stochastic nonlinear systems subject to command filter and time-delay interactions. The main features are (1) the communication interconnection has the output delay between subsystems; (2) the smooth finite time strategy is given for stochastic disturbances; (3) all states are confined in the constrained space without feasibility conditions. First, the state time-delay interactions are tackled by the appropriate establishment of Lyapunov-Krasovskii functionals. Second, for full-state constraints, state-related mappings are employed to remove the feasibility. Then, by constructing the switching virtual control signals, the smooth fast finite-time feature is attached to the controller. Finally, based on finite-time stability theory in probability, command filter, and backstepping technique, the smooth fast finite-time neural controller is constructed to make that all system states converge in finite time, all closed-loop variables are bounded in probability, and full-state constraints are satisfied. Simulation results are provided to test the correctness of the method.