基于UKF的水下机器人执行器故障检测方法研究  

Research on actuator fault detection of unmanned underwater vehicle based on unscented kalman filter

作　　者：林昌龙[1,2] 封锡盛[1] 李一平[1,2] 

机构地区：[1]中国科学院沈阳自动化研究所机器人学国家重点实验室,沈阳110016 [2]中国科学院研究生院,北京100049

出　　处：《机械设计与制造》2011年第5期168-170,共3页Machinery Design & Manufacture

基　　金：国家自然科学基金(60775061);国家基础研究项目(6138102005-4;6138101004-3)

摘　　要：水下机器人需要在无人干预或者少量干预的情况下自主地完成使命,这就要求它在作业过程中能够自主地检测子系统、传感器和执行器的故障。在确认发生故障的时候,它还需要尽最大可能地进行修复,以确保使命的顺利进行。水下机器人常用的执行器包括推进器、舵、机械手等,针对推进器的故障检测问题,首先建立了水下机器人的运动学和动力学模型,然后采用UKF对系统的状态和参数进行联合估计,最后通过外场试验验证了算法的有效性和正确性。Unmanned underwater vehicle needs to complete the mission with little or even without human intervention.Thus,it should be able to detect the faults of its subsystems,sensors,and actuators during the mission.Furthermore,when identifying any fault,it should try to fix the problem to resume the mission.Generally,the actuators for unmanned underwater vehicles consist of thrusters,rudders,manipulators and so on.Concerning the problem of actuator faults detecting,the kinematical and kinetic models of an unmanned underwater vehicle are firstly constructed it.Then using an Unscented Kalman Filter,the status and parameters of the system are estimated.Finally,field experiment is carried out and the result validates the effectiveness and correctness of this algorithm.

关 键 词：故障检测 水下机器人 无色卡尔曼滤波器 

分 类 号：TH16[机械工程—机械制造及自动化] TP242.3[自动化与计算机技术—检测技术与自动化装置]