计算机科学
控制理论(社会学)
反推
模糊逻辑
控制器(灌溉)
容错
执行机构
人工神经网络
断层(地质)
跟踪误差
小波
自适应控制
人工智能
控制(管理)
分布式计算
地震学
农学
生物
地质学
作者
Xiaona Song,Chenglin Wu,Shuai Song,Vladimir Stojanović,Inés Tejado
标识
DOI:10.1016/j.engappai.2023.107832
摘要
This paper focuses on fuzzy wavelet neural networks-based adaptive finite-time self-triggered fault-tolerant control design with assured tracking performance for a quadrotor unmanned aerial vehicle subject to unknown actuator faults. First, an improved finite-time performance function is integrated with the command-filtered backstepping control framework to assure the scheduled tracking performance, which can effectively constrain the transient fluctuations of the tracking error at fault occurrence. Then, two compensation mechanisms are developed to weaken the adverse impact induced by actuator faults and filter errors. Further, a fuzzy wavelet neural adaptive self-triggered fault-tolerant controller with guaranteed performance is designed to improve the fault insensitivity and tracking accuracy of the controlled vehicle, where the fuzzy wavelet neural networks are employed to approximate the unknown nonlinearities and the control signals are allowed to achieve irregular updating only when the pre-specified trigger protocol is violated. Stability analysis proves that the designed controller guarantees the attitude and position subsystems are practical finite-time stable, and the tracking errors are strictly confined to a preassigned region and never cross its allowed bounds. Finally, the validity of the developed control scheme is confirmed by the simulation results.
科研通智能强力驱动
Strongly Powered by AbleSci AI