CoLRIO is a system for collaborative state estimation using LiDAR-ranging-inertial data in robotic swarms operating in GPS-denied environments. It centralizes computationally intensive tasks to a server, reducing the computational burden on individual robots. The server refines joint pose graph optimization through place recognition, global optimization techniques, and removal of outlier data. The system improves the accuracy of collaborative SLAM estimates and proves effective in large-scale missions. The research team will make their code open-source and accessible.
Publication date: 20 Feb 2024
Project Page: https://github.com/PengYu-team/Co-LRIO
Paper: https://arxiv.org/pdf/2402.11790
