摘要
针对复杂海洋环境中水下机器人的可靠性问题,提出了一种水下机器人的主动容错控制方法.在分析六自由度水下机器人受力的基础上建立了水下机器人简化动力学模型;利用快速学习的递归小脑神经网络(RCMAC)学习辨识机器人中出现的时变故障,并根据故障辨识结果重新配置控制律,使无人水下机器人在故障情况下仍然可以完成预定任务.仿真实验证明了该容错方法的有效性.
An active fault-tolerant control method is proposed to ensure the reliability of the underwater vehicle in the complicated ocean surroundings.Firstly,the simplified dynamic model is obtained for the 6-DOF(degree of freedom) underwater vehicle based on the analysis of its dynamics.Then the recurrent cerebellar model articulation controller(RCMAC) is proposed to identify the time-varying faults for such 6-DOF underwater vehicle.The control law is reconfigured according to the result of the fault identification for the purpose of making the underwater vehicle accomplish the scheduled mission while fault occurs.The proposed identification and control strategy is illustrated by a simulation with its efficiency being shown.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2009年第S1期147-150,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(50775136)
国家高技术研究发展计划资助项目(2006AA09Z210)
上海市教委支出预算项目(2008099)
关键词
水下机器人
故障辨识
神经网络
递归小脑神经网络
主动容错控制
underwater vehicle
fault identification
neural network
recurrent cerebellar model articulation controller
active fault-tolerant control
