故障检测与隔离
稳健性(进化)
观察员(物理)
控制理论(社会学)
计算机科学
容错
计算
国家观察员
断层(地质)
实时计算
控制工程
分布式计算
工程类
控制(管理)
算法
人工智能
执行机构
非线性系统
地质学
物理
地震学
基因
化学
量子力学
生物化学
作者
Li‐Guo Qin,Xiao He,Rui Yan,Ruiliang Deng,Donghua Zhou
标识
DOI:10.1016/j.jfranklin.2017.11.020
摘要
In the paper, a distributed sensor fault detection and isolation scheme is presented for a network of second-order integrators. A new distributed control law is developed to achieve formation of the system. By using the integration information of distributed formation errors, the control law improves the robustness of the formation. A distributed observer is then designed in each vehicle based on the closed-loop dynamic model of the vehicle. Each vehicle updates the states of the distributed observer by employing the measurements of itself and the transmitted state estimations from its neighbors. Based on the distributed observer, a distributed fault detection observer and a distributed fault isolation observer are designed. The presented distributed fault detection observer in each vehicle is able to be sensitive to the faults of all vehicles in the system. By using the distributed fault isolation observers, each vehicle is able to be sensitive to the faults of itself, its neighbors and its neighbors’ neighbors and to be robust to the faults of other vehicles. Although the fault isolation of the proposed scheme is simple, computation loads of the scheme are lower than the current ones since only the model of the individual vehicle is used. Finally, the effectiveness of the control law and the fault diagnosis scheme is demonstrated by simulations and real-time experiments carried out based on a formation of three quadrotors.
科研通智能强力驱动
Strongly Powered by AbleSci AI