【摘要】 为提高遥控水下机器人的可靠性,研究了推力分配建模和基于递归神经网络的容错控制技术.以SY-Ⅱ遥控水下机器人为研究对象,针对其推力器布置情况进行推力分配建模,并以该模型为基础通过对偶原理建立机器人控制的能量函数,实现了递归神经网络形式的推力器容错控制算法.通过容错控制的仿真和实验表明,该算法具有较强的鲁棒性,能够根据推力器的不同故障状况,重新进行推力分配.使得要遥控水下机器人当一个主推和其它任意一个平面推力器正常工作,而其它平面推力器发生故障时,仍能准确地实现机器人的水平面两自由度容错运动控制.从而说明该算法在容错控制方面的优越性,提高遥控水下机器人的安全性和可靠性.更多还原

【Abstract】 In order to improve the reliability of Remote Operated Vehicles(ROV),force allocation model and fault tolerant control based on recurrent neural network were studied.In the research of SY-Ⅱ ROV thrusters’ position,force allocation model had been established.On the Basis of this model,the energy function concerning ROV control had been established through duality priciple.Thus the fault tolerant algorithm with recurrent neural network format had been developed.Simulations and experiments results displayed that algorithm was very robust and could reallocate thrust force according to different fault conditions.The horizontal 2-DOF fault tolerant control can be accomplished when one main thruster and any other one working properly,while others in fault.Therefore the algorithm is proved superiority,the reliability and safety for ROV have been improved.更多还原