每天一分钟，带你读遍机器人顶级会议文章
标题：DOOR-SLAM: Distributed, Online, and Outlier Resilient SLAMfor Robotic Teams
作者：Pierre-Yves Lajoie,Benjamin Ramtoula
来源：2020 IEEE International Conference on Robotics and Automation (ICRA)
编译：张宁
审核：柴毅，王靖淇
这是泡泡一分钟推送的第 659 篇文章，欢迎个人转发朋友圈；其他机构或自媒体如需转载，后台留言申请授权
摘要
为了完成协作任务，团队中的机器人需要对环境及其在环境中的位置有共同的了解。分布式同时定位和地图绘制（SLAM）提供了一种实用的解决方案，可以在不依赖外部定位系统（例如GPS）且信息交换最少的情况下对机器人进行定位。不幸的是，当前的分布式SLAM系统容易受到感知异常值的影响，因此倾向于使用非常保守的参数进行机器人间位置识别。但是，过于保守会以拒绝许多有效的闭环候选者为代价，这会导致轨迹估计的准确性降低。本文介绍了DOOR-SLAM，这是一个完全分布式的SLAM系统，具有离群剔除机制，可以使用较少的保守参数。DOOR-SLAM基于对等通信，不需要机器人之间的完全连接。DOOR-SLAM包含两个关键模块：姿态图优化器与分布式成对一致的测量集最大化算法相结合，以拒绝伪造的机器人间环路闭合；以及分布式SLAM前端，可检测机器人之间的回路闭合而无需交换原始传感器数据。该系统已在仿真，基准数据集和野外实验（包括在GPS受限的地下环境中进行的测试）中进行了评估。DOOR-SLAM产生更多的机器人间回路闭合，成功地排除异常值，并在不需要低通信带宽的情况下得出准确的轨迹估计。完整的源代码可从https://github.com/MISTLab/DOOR-SLAM.git获得。
图1：DOOR-SLAM（红色和蓝色）和GPS地面真实情况（绿色，仅用于基准测试）的轨迹估计。
图2：DOOR-SLAM系统概述
图3：分布式闭环检测概述。
Abstract
To achieve collaborative tasks, robots in a team need to have a shared understanding of the environment and their location within it. Distributed Simultaneous Localization and Mapping (SLAM) offers a practical solution to localize the robots without relying on an external positioning system (e.g. GPS) and with minimal information exchange. Unfortunately, current distributed SLAM systems are vulnerable to perception outliers and therefore tend to use very conservative parameters for inter-robot place recognition. However, being too conservative comes at the cost of rejecting many valid loop closure candidates, which results in less accurate trajectory estimates. This paper introduces DOOR-SLAM, a fully distributed SLAM system with an outlier rejection mechanism that can work with less conservative parameters. DOOR-SLAM is based on peerto-peer communication and does not require full connectivity among the robots. DOOR-SLAM includes two key modules: a pose graph optimizer combined with a distributed pairwise consistent measurement set maximization algorithm to reject spurious inter-robot loop closures; and a distributed SLAM front-end that detects inter-robot loop closures without exchanging raw sensor data. The system has been evaluated in simulations, benchmarking datasets, and field experiments, including tests in GPS-denied subterranean environments. DOOR-SLAM produces more inter-robot loop closures, successfully rejects outliers, and results in accurate trajectory estimates, while requiring low communication bandwidth. Full source code is available at https://github.com/MISTLab/DOOR-SLAM.git.