灌溉机器人全覆盖路径规划方法  

Full Coverage Path Planning Method for Irrigation Robots

作　　者：王臻卓[1] 陈金林 任婷婷[2] 杨科科[1] 任宁宁 WANG Zhen-zhuo;CHEN Jin-lin;REN Ting-ting;YANG Ke-ke;REN Ning-ning(Department of Automation Engineering,Henan Polytechnic Institute,Nanyang 473000,Henan Province,China;School of Electric Power and Architecture,Shanxi University,Taiyuan 030000,China;School of Electrical Engineering and Automation,Henan Polytechnic University,Jiaozuo 454003,Henan Province,China)

机构地区：[1]河南工业职业技术学院自动化工程学院,河南南阳473000 [2]山西大学电力与建筑学院,太原030000 [3]河南理工大学电气工程与自动化学院,河南焦作454003

出　　处：《节水灌溉》2024年第9期53-58,共6页Water Saving Irrigation

基　　金：河南省高校重点科研项目(24B470003);河南省高校重点科研项目(22B470003)。

摘　　要：灌溉机器人全覆盖行动的各个任务具有较为明显的空间并行性,随着全覆盖范围扩大,在对覆盖区域进行分解阶段,需要充分考虑将整个区域空间分解为哪些区域。但是,灌溉机器人受到视觉感知区域限制,准确匹配和衔接路块间最近端点的难度较大,导致局部路点的连通和线路衔接出现差错,难以有效全覆盖。为了有效解决这一问题,提出一种灌溉机器人全覆盖路径规划方法。通过快速搜索随机算法展开需要覆盖区域的边界检测,考虑视觉传感器的感知范围受限因素,采用灰度质心法展开区域视图边界提取,根据提取结果建立地图。在地图上建立线段序列,通过曼哈顿最小距离原则连接地图上的部分路径线段,形成多个弓形线路块。使用分治算法匹配和衔接各个弓形线路块间最近端点对,引入改进A*算法对全局以及局部路点的连通和线路衔接,实现灌溉机器人的全覆盖路径规划。实验结果表明:针对简单灌溉区域,该方法的路径重复率为0.041%,灌溉覆盖率为98.90%;针对复杂灌溉区域,该方法的路径重复率为0.017%,灌溉覆盖率为99.87%。这说明针对不同的灌溉环境,该方法均可以实现理想的路径规划,不仅可以最大限度地实现全覆盖,并有效地减少路径冗余程度,可以获取理想的灌溉机器人全覆盖路径规划方案。The tasks of the irrigation robot′s full coverage operation have obvious spatial parallelism.As the full coverage range expands,during the decomposition stage of the coverage area,it is necessary to fully consider how to decompose the entire area into subareas.However,irrigation robots are limited by their visual perception area,making it difficult to accurately match and connect the nearest endpoints between path segments,resulting in errors in the connection of local road points and lines,making it difficult to effectively achieve full coverage.To address this issue effectively,a full coverage path planning method for irrigation robots is proposed.By using a fast search random algorithm to expand the boundary detection of the target area,considering the limited perception range of visual sensors,the grayscale centroid method is used to perform boundary extraction of the area view,and a map is established based on the extraction results.A sequence of line segments is established on the map,connecting some path segments using the Manhattan minimum distance principle,and form multiple arched line blocks.Using the divide and conquer algorithm to match and connect the nearest endpoint pairs between various bow shaped line blocks,in an improved A*algorithm is introduced for global and local road point connectivity and line connection,achieving full coverage path planning for the irrigation robots.The experimental results show that for simple irrigation areas,the path repetition rate of this method is 0.041%,and the irrigation coverage rate is 98.90%;For complex irrigation areas,the path repetition rate of this method is 0.017%,and the irrigation coverage rate is 99.87%.This indicates that for different irrigation environments,this method can achieve ideal path planning,not only maximizing full coverage,but also effectively reducing path redundancy,and obtaining an ideal full coverage path planning scheme for irrigation robots.

关 键 词：灌溉机器人 全覆盖线路 路径规划 快速搜索随机算法 边界提取 分治算法 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置] S275[自动化与计算机技术—控制科学与工程]