基于均匀扫描和专注引导策略的自主探索算法  

Autonomous exploration algorithm based on uniform scanning and attentive guidance explorer

作　　者：申伟霖 陈荟慧 关柏良 王爱国 杨健茂 Shen Weilin;Chen Huihui;Guan Boliang;Wang Aiguo;Yang Jianmao(School of Electromechanical Engineering&Automation,Foshan University,Foshan Guangdong 528225,China;School of Electronic&Information Engineering,Foshan University,Foshan Guangdong 528225,China)

机构地区：[1]佛山科学技术学院机电工程与自动化学院,广东佛山528225 [2]佛山科学技术学院电子信息工程学院,广东佛山528225

出　　处：《计算机应用研究》2024年第11期3415-3419,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(61972092);广东省教育厅重点领域专项资助项目(2022ZDZX1026)。

摘　　要：为解决机器人难以快速探索含有狭窄入口的未知环境,以及在探索收益接近的目标之间徘徊探索而导致探索效率降低的问题,提出一种包含均匀扫描和专注引导策略的自主探索算法USAGE。USAGE采用均匀扫描的方式检测地图的边界点,并对边界点进行聚类得到待探索点。最后通过专注引导策略确定最优探索目标,在含有信息增益和路径代价的传统评价函数中引入转向代价评估指标,并根据机器人的状态约束探索任务执行,引导机器人专注探索。通过在机器人操作系统搭建仿真环境进行验证,实验结果表明,与基于快速探索随机树(rapidly-exploring random tree, RRT)的探索算法相比,USAGE占用系统内存减少了11.34%以上,在探索耗时和探索距离方面分别减小了26.90%和31.94%,提升了自主探索效率。To solve the problem of difficulty for robots to quickly explore unknown environments with narrow entrances,as well as the decrease in exploration efficiency caused by wandering between targets with close exploration benefits,this paper proposed a self-exploration algorithm USAGE(uniformly scanning and attentive guidance explorer),which included uniform scanning and attentive guidance strategy.USAGE used the uniform scanning method to detect boundary points,and clustered the boundary points to obtain the points to be explored,and finally determined the optimal exploration target through an attentive guidance strategy.This strategy introduced the steering cost evaluation index into the traditional evaluation function containing information gain and path cost,and constrained the execution of the exploration task according to the robot’s state,and guided the robot to focus on exploration.Through building a simulation environment in the robot operating system for verification,the experimental results show that compared with the exploration algorithm based on rapidly-exploring random tree(RRT),USAGE occupies less system memory by more than 11.34%,reduces the exploration time and exploration path level by 26.90%and 31.94%respectively,which improves autonomous exploration efficiency.

关 键 词：机器人 自主探索 边界点 评价函数 机器人操作系统 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]