多机器人协同SLAM技术研究进展

由于单机器人同步定位与建图（SLAM）技术在实际应用中的局限性，多机器人协同SLAM技术以较强的灵活性和鲁棒性受到研究人员的广泛关注，并且在农业生产、环境监测、海上搜救等领域具有巨大应用前景。多机器人协同SLAM是多机器人协同工作的核心及大范围复杂环境内及时获得场景感知信息的关键，能使多个机器人在协同工作时共同定位并构建任务空间地图，主要基于单机器人SLAM算法、多机器人系统架构、地图融合等技术实现。结合多机器人协同SLAM的发展历程，对比分析当前主流的多机器人协同SLAM算法。从传感器的角度，将多机器人协同SLAM分为激光协同SLAM、视觉协同SLAM以及激光视觉融合协同SLAM三类，并对多机器人协同SLAM的架构选择、多机通信、相对位姿、地图融合和后端优化问题进行讨论，同时指出异构机器人协同、基于深度学习的语义SLAM是多机器人协同SLAM的未来发展趋势。

Research Progress of Multi-Robot Collaborative SLAM Technology

Because of the limitations of single-robot Simultaneous Localization and Mapping(SLAM) technology in practical applications, multi-robot collaborative SLAM technology has received extensive attention from researchers with its strong flexibility and robustness, and has great application prospects in agricultural production, environmental monitoring, maritime search and rescue, and other fields.Multi-robot collaborative SLAM is the core of multi-robot collaborative work, and it is the key to obtaining timely situational awareness information in a large-scale complex environment, which enables each robot to co-locate and build a workspace map when working together.Multi-robot collaborative SLAM is mainly implemented based on single-robot SLAM algorithm, multi-robot system architecture, map fusion and other technologies.Combined with the development history of multi-robot collaborative SLAM, this study compares and analyzes the current mainstream multi-robot collaborative SLAM algorithms.From the sensor, the multi-robot collaborative SLAM is divided into three categories:laser collaborative SLAM, vision collaborative SLAM and laser vision fusion collaborative SLAM, and the architecture selection, multi-machine communication, relative pose, map fusion and post-processing of multi-robot collaborative SLAM are discussed.Simultaneously, it is pointed out that the collaborative SLAM of heterogeneous robots and semantic SLAM based on deep learning is the future development trend of multi-robot collaborative SLAM.