摘要
考虑博物馆内存在的静态障碍物和动态障碍物,为保证博物馆机器人可动态调整路径以避开多种障碍物,自主运行至目标点,研究基于改进机器学习的博物馆机器人自主避障控制方法。该方法构建机器人工作环境栅格地图,并通过改进Q-learning算法的奖励函数,使机器人能够探索出抵达目标点最短且能避开障碍的路径。利用模糊PID控制器设计模糊轨迹控制器,根据机器人当前位置与规划路径之间的航向角误差及其变化率,动态调节转向角控制量,使机器人能沿规划路径自主运行。实验结果表明,在离散型与聚集式混合型障碍物场景中,该方法控制下的机器人轨迹控制误差最大值分别为1.46°和1.21°,且碰撞次数为零,验证了该方法可为博物馆机器人规划距离较短且安全的自主避障路径。
Considering the static and dynamic obstacles within museums,this study investigates an autonomous obstacle avoidance control method for museum robots based on improved machine learning.The approach constructs a grid map of the robot’s working environment and employs an enhanced Q-learning algorithm with a reward function to enable the robot to explore the shortest path to the target while avoiding obstacles.A fuzzy PID controller is designed to create a fuzzy trajectory controller,which dynamically adjusts the steering angle control based on the heading angle error between the robot’s current position and the planned path,as well as its rate of change,allowing the robot to autonomously follow the planned path.Experimental results show that in scenarios with mixed discrete and clustered obstacles,the maximum trajectory control errors under this method are 1.46°and 1.21°,respectively,with zero collisions,verifying that the method can plan shorter and safer autonomous obstacle avoidance paths for museum robots.
作者
周娜
ZHOU Na(Shandong Sport University,Jinan 250000,China)
出处
《电子设计工程》
2026年第3期83-88,共6页
Electronic Design Engineering
关键词
机器学习
机器人自主避障
栅格地图
模糊轨迹控制器
machine learning
robot autonomous obstacle avoidance
grid map
fuzzy trajectory controller
