基于神经网络的滑模控制在水下机器人中的应用  

Application of Sliding Mode Control Based on Neural Networks to Underwater Robot

作　　者：李敏[1] 刘和平[2] 陈永刚[1] 刘寻[1] 

机构地区：[1]河南科技大学电子信息工程学院,河南洛阳471003 [2]洛阳理工学院机电工程系,河南洛阳471023

出　　处：《河南科技大学学报（自然科学版）》2010年第3期18-21,共4页Journal of Henan University of Science And Technology:Natural Science

基　　金：国家自然科学基金项目(40771166)

摘　　要：在水下机器人的控制中,对强耦合非线性运动进行精确建模和解析是极其困难的,本文采用神经网络对水下机器人水动力引起的不确定项和干扰的上限进行自适应逼近学习,然后采用滑模变结构方法对水下机器人航向进行控制,并进行了跟踪仿真和静态水池实验,结果表明:该种控制方法具有一定的效果。In the control of underwater robot,it is difficult to establish and resolve the precision dynamic model.This paper presents a variable structure sliding mode control for underwater robot,with which the adaptive learning of the neural network is used to estimate and approach the upper bound of the uncertainty and disturbance induced by hydrodynamics.According to the description and setting up of the control model,a tracking simulation is carried out and a series of tests on the yaw of underwater robot are performed in static pool.It is proved that this control strategy is available for the underwater robot.

关 键 词：神经网络 滑模控制 水下机器人 自适应学习 

分 类 号：U674.941[交通运输工程—船舶及航道工程] TP389.1[交通运输工程—船舶与海洋工程]