摘要
该文设计实现了一种针对集群机器人协同定位的实验平台。该平台旨在通过模拟真实环境中的复杂场景,为研究集群机器人如何在给定的分组结构内高效协同工作以实现精确的自我定位,提供可靠的测试与验证工具。本设计使用超宽带传感器、惯性测量单元传感器以及动作捕捉系统实现对集群机器人的运动控制,无需采用全球定位系统、激光雷达等高精度设备,不仅压缩了集群机器人协同定位的传感器成本,而且极大降低了任务的算力消耗。该平台还综合考量了系统的可操作性和监控需求,设计并实现了一套上位机系统。该实验平台融合了最新的机器人技术,为学生提供了一个集学习、研究于一体的平台,可应用于高校自动化、人工智能等相关专业的教学与实践课程。
[Objective]Swarm robotics technology has exhibited significant potential in high-precision localization tasks,emerging as a cutting-edge research focus in automation and intelligent systems.As localization tasks grow increasingly complex,the required scale of swarm robotics has surged,and the use of large-scale robotic swarms to efficiently implement complex cooperative localization has become critically important.To achieve this,the grouping constraints strategy is applied to divide large-scale robotic swarms into multiple small-scale subgroups,each comprising two to five robots.Within each subgroup,one robot is designated as the central robot,whereas the other robots are the member robots.The central robot is responsible for information exchange and coordinated motion with the central robots from the other subgroups,as well as for task decomposition and allocation to member robots within its subgroup.This hierarchical structure significantly streamlines the roles and responsibilities of swarm robots and enables the entire swarm to implement complex localization tasks more cohesively and efficiently.Furthermore,the communication mechanisms among robots are significantly simplified,hence eliminating the risks of communication interference and conflicts.[Methods]Motivated by these advancements,this study designs an experimental platform for swarm robotics cooperative localization based on grouping constraints.The platform simulates complex real-world scenarios to investigate how swarm robots efficiently collaborate within predefined group structures,providing a reliable testing and validation tool for precise localization in large-scale robotic swarms.By utilizing a hierarchical management and distributed execution framework,the platform decomposes complex system-wide motion tasks into cluster-level operations.Intercluster motion control includes fundamental modes,such as separation,convergence,and halt.The central robots interpret cluster-level tasks assigned by the host computer,plan trajectories,and allocate subtasks to the member robots within their subgroups.Intra-cluster motion control encompasses basic modes,such as forward/backward movement,lateral translation,and circular motion.Upon receiving orders from the central robot,the member robots execute the corresponding actions based on their capabilities and real-time environmental feedback.To avoid high-cost devices,such as GPS or LiDAR,the platform employs customized smart vehicles equipped with ultra wide band(UWB)sensors,inertial measurement units,and a motion capture system for motion control and performance evaluation.This configuration not only reduces sensor costs but also minimizes computational overhead for swarm robotics cooperative localization.In addition,the integration of the host computer with on-site cameras enables real-time monitoring of swarm behavior through a dedicated interface,facilitating dynamic strategy adjustments and rapid fault diagnosis.[Results]To validate the availability of the designed swarm robotics cooperative localization platform,the UWB sensors were used to acquire the relative distance between robots,whereas the Vicon motion capture system was utilized to obtain the ground truth positions of the robots.Three scenarios,namely,fixed-distance swarm movement,point-to-point swarm navigation,and circular swarm motion,were analyzed,and the experimental results showed that:(1)When the robots moved with the fixed distance of 1 m,the minimum average localization error was-0.06233 m,every effective result compared with 1 m.In addition,the localization accuracy decreased as the distance between UWB sensors deployed at the four vertices of the central robot increased.(2)When the member robots were required to navigate to the fixed point(1 m,1 m),the lowest average cooperative localization error was 0.03973 m,indicating accurate cooperative localization.(3)When the member robots orbited the central robot at a radius of 2 m,the localization error increased to 1.96014 m because of the escalated UWB positioning error when the robots were operating in a complex circular motion.[Conclusions]By integrating state-of-the-art swarm robotics technology based on grouping constraints,this system can serve as a multifunctional educational platform,enabling students to engage in both theoretical learning and hands-on research and can be directly applicable to teaching and practical courses in automation,artificial intelligence,and related disciplines at the university level,bridging the gap between academic concepts and real-world swarm robotics applications.
作者
石义芳
李真
叶沛欣
SHI Yifang;LI Zhen;YE Peixin(College of Automation,Hangzhou Dianzi University,Hangzhou 310018,China)
出处
《实验技术与管理》
北大核心
2025年第5期54-62,共9页
Experimental Technology and Management
基金
国家自然科学基金项目(62471165)。
关键词
集群机器人
分组约束
协同定位
超宽带传感器
惯性测量单元
swarm robotics
grouping constraints
cooperative localization
ultra wide band(UWB)
inertial measurement unit(IMU)
