基于神经网络自适应的水下移动机器人轨迹跟踪方法

打开文本图片集

中图分类号：TP242.3 文献标志码：B 文章编号：1671-5276（2025）05-0219-05

Adaptive Trajectory Tracking Method for Underwater Mobile Robots Based on Neural Networks

ZHENG Jun

（Yantai Salvage Bureau of the Ministryof Transport，Yantai 264Ol2，China）

Abstract：Basedontheadaptivealgorithmof neural network，atrajectory tracking methodis designedaccording tothe workingenvironment of underwater mobile robot.Onthe basisof the motion dimension of the underwater robot，the coresponding parameters areset，andthethree-dimensional kinematics is converted intoamodel.The Bernoulli Newton method is used to model the current and wave，and the motion model of the underwater environment robot is obtained.The tracking data is saved in the form ofconstant neural network weights，and the output data of the neural network are online learned.Theadaptive learning controlleris designedtochange the internal parameters of the controler，therobot motionstatetrackingcontrolisrealized，andthefedback controlisdynamicallyrealizedaccording tothe trajectory tracking eror.The simulation modelof mobilerobotis built in Simulink environment，and diferent trajectorytracking methods areused to testand compare the performance.Thefinal results show that under diferent workingconditions，the proposed methodcanreduce theerorof trajectory tracking andeffectivelyrealizetheacurate trajectory tracking of underwater mobile robot.

Keywords：underwater mobile robots；neural networks；adaptive；trajectory tracking

0 引言

避开障碍物，减少碰撞风险，从而确保机器人的安全。（剩余5215字）