摘要
随着无人配送的推广与应用,自动导引运输车(AGV)的导航方法成为了研究的热点。主要针对校园配送AGV的地图构建和路径规划展开研究,地图构建由即时定位与地图构建(SLAM)来实现,路径规划则根据已构建好的地图,在点与点之间计算好最优路线。采用基于LOAM的算法,将SLAM问题分为高频的运动估计和低频的环境建图。在运动估计中通过曲率计算提取平面点和边缘点两种特征,采用scan-to-scan方式进行特征匹配,并建立误差模型。在环境建图中采用map-to-map方式匹配,运用PCA算法提取点云中的特征点,利用LM算法来求解。再结合A*算法和DWA算法获得最优路径,并提高了对环境的适应能力。实验结果表明,提供的地图构建及路径规划方法能够满足AGV在校园环境多种不同模拟场景中的定位导航需求,具有一定的可行性和可靠性。
With the popularization and application of unmanned delivery, the method of AGV navigation has become a research hot spot.Focusing on the mapping and path planning of the AGV for campus delivery, the mapping is realized by SLAM, and the path planning calculates the optimal route between points according to the mapping. Using the algorithm based on LOAM, the SLAM problem was divided into high-frequency motion estimation and low-frequency environment mapping. In motion estimation, two kinds of features, plane point and edge point, were extracted by curvature calculation, and the scan to scan method was used for feature matching, and then the error model was established. In the environmental mapping, map to map matching was adopted, PCA algorithm was used to extract the feature points in the point cloud, and LM algorithm was used to solve the problem. Combined with A* algorithm and DWA algorithm, the optimal path was obtained,and the adaptability to the environment was improved. The experimental results show that the mapping and path planning methods provided can meet the AGV positioning and navigation requirements in a variety of different simulation scenes in the campus environment, and have certain feasibility and reliability.
作者
朱伟枝
谢娟烘
卢子荣
Zhu Weizhi;Xie Juanhong;Lu Zirong(Guangdong Technology College,Zhaoqing,Guangdong 526100,China)
出处
《机电工程技术》
2023年第1期107-110,160,共5页
Mechanical & Electrical Engineering Technology
基金
广东理工学院科技项目(编号:2021GKJZK012)。
