室内多机器人协作搜索方法研究

【作者】 杨帆；

【导师】 李胜；

【作者基本信息】 南京理工大学 ， 控制理论与控制工程， 2021， 硕士

【摘要】 近年来多机器人系统作为机器人研究领域的一部分吸引了广大研究者的关注,在军事侦察、资源探测、危险环境巡检救援等多种场合得到了非常广泛的应用。多机器人协作目标搜索更是多机器人研究中的热点方向。本文针对室内未知环境下多机器人协作搜索存在的搜索效率低、时间长等问题,利用群体智能算法,研究了多机器人协作搜索问题,完成的主要工作如下:首先,阐述了多机器人系统的研究背景和意义,分析多机器人协作搜索相关技术国内外研究现状,及其存在的问题和面临的挑战。建立了多机器人目标搜索问题的数学模型,分析了粒子群算法基本思想和多机器人协作搜索相关概念的相关性,及将粒子群算法运用于多机器人协作搜索中的合理性。然后,针对室内未知环境下静态单目标搜索问题,提出了一种基于粒子群优化与天牛须优化算法的多群融合算法。该算法借鉴多群融合思想,扩大了多机器人系统的搜索范围,提高了多机器人协作搜索的成功率。并通过仿真验证了,该算法在目标搜索成功率和搜索效率方面优于传统目标搜索算法。接着,针对环境对目标搜索成功率影响的问题,提出了一种基于目标概率地图的多机器人协作搜索方法。该算法利用环境中的目标信号模型与机器人获取的目标信号信息建立目标概率地图,通过所设计的多机器人协作搜索策略,选择下一时刻运动的目标点,同时在室内大范围障碍物的情况下(如墙壁等),使用D*Lite算法到达搜索策略选择的目标点,降低了环境对目标搜索成功率的影响。最后,搭建了室内多机器人协作搜索实验平台。该平台基于ROS2的Ardent框架,提高了多机器人协作的可靠性,借助已有的机器人导航定位功能包,对本文所提出的多机器人协作搜索算法进行了测试和验证。实验结果表明,利用基于目标概率地图的多机器人协作搜索算法,多机器人可以快速准确地搜索到目标。更多还原

【Abstract】 In recent years,the attention of researchers have been attracted to multi-robot system which is a part of the field of robotics research and has been widely used in military reconnaissance,resource detection,and hazardous environment inspections and rescues.Multi-robot collaborative target search is a hot topic in multi-robot research.Aiming at the problem of the efficiency and time of multi-robot target search in an unknown indoor environment,swarm intelligence algorithms is used to solve the problem of multi-robot collaborative search.The main tasks completed are as follows:First,the research background and significance of the multi-robot system are explained,and the domestic and foreign research status of the multi-robot collaborative search related technology is analyzed,as well as the existing problems and challenges.The mathematical model of the multi-robot target search problem is established,the basic idea of the particle swarm algorithm and the related concepts of the multi-robot cooperative search are analyzed,and the rationality of applying the particle swarm algorithm to the multi-robot cooperative search is analyzed.Then,aiming at the static single-object search problem in an unknown indoor environment,a multi-group fusion algorithm based on particle swarm optimization and longhorn beetle optimization algorithm is proposed.The algorithm draws on the idea of multi-group fusion,expands the search range of the multi-robot system,and improves the success rate of multi-robot collaborative search.It is verified by simulation that the algorithm is superior to traditional target search algorithms in terms of target search success rate and search efficiency.Then,in view of the influence of environment on target search success rate,a multi-robot cooperative search method based on target probability map is proposed.The algorithm uses the target signal model in the environment and the target signal information obtained by the robot to establish a target probability map.Through the designed multi-robot collaborative search strategy,the target point to move at the next moment is selected,and at the same time in the case of a large range of indoor obstacles(Such as walls,etc.),use the D*Lite algorithm to reach the target point selected by the search strategy,which reduces the influence of the environment on the success rate of target search.Finally,an indoor multi-robot collaborative search experiment platform is built.The platform is based on the Ardent framework of ROS2,which improves the reliability of multi-robot collaboration.With the help of the existing robot navigation and positioning function package,the multi-robot collaborative search algorithm proposed in this paper is tested and verified.The experimental results show that using the multi-robot cooperative search algorithm based on the target probability map,multi-robots can quickly and accurately search for the target.更多还原