Collaborative Exploration and Mapping with Multimodal LiDAR Sensors

Exploration and mapping unknown environments have placed elevated demands on the spatial intelligence perception capabilities of unmanned systems. Traditional single unmanned systems suffer from issues such as low exploration efficiency. This paper introduces the LuoJia-Explorer system, an unmanned multi-robot system that can achieve autonomous collaborative exploration and mapping with multimodal LiDAR sensors. For individual agent in the system, the robot achieves tightly-coupled multi-LiDAR-inertial odometry and mapping, thereby leveraging this capability to autonomously explore unknown spaces by computing visibility maps and determining the shortest collision-free paths. On the multi-unmanned system level, the system combines the local poses and maps of individual robots to accomplish collaborative multi-robot localization and mapping. Moreover, the robots within the system exchange information regarding point cloud submaps and relative spatial exploration status, allowing for the sharing of field of view. This collaborative framework enables autonomous path planning among multiple robots. We have conducted practical validation to assess the effectiveness of the proposed system on the campus of Wuhan university. The system achieves an average point-to-point distance of 0.12m between the acquired and ground truth maps that are produced by high-precision TLS instruments. Additionally, the system exhibits a spatial exploration efficiency of 478.104 m3/s in the test scene. Our system has achieved an 86.38% increase in exploration efficiency compared to a single robot, and our system's spatial coverage rate can reach 86.70%.