Introducing IRL SLAM Studio

Lab News ·

Description

At the Intelligent Robots Lab (IRL), Chungbuk National University, we’ve been building a SLAM system that doesn’t care what it’s mounted on or where it’s driving. Today we’re sharing our introduction video.

What IRL SLAM Studio does:🤖Runs on wheeled robots, quadrupeds, UGVs and agricultural tractors🌲Works indoors, on campus, in forests and in open fields, including feature-poor scenes and rough terrain🔌Plug-and-play sensors: LiDAR (Velodyne, Hesai, Ouster, Livox), RGB-D cameras (RealSense, ZED) and various IMUs⚙️Flexible fusion modes: VIO, LIO, LVIO, with optional GNSS, plus multi-LiDAR, multi-camera and auto-calibration📡100+ Hz pose output on ROS and ROS 2💻Lightweight enough for NVIDIA Jetson OrinReal-world results, with no loop closure and no GPS fusion:•Indoor corridor: 2.7 cm after 106 m•Campus: 7.2 cm over 458 m•Forest slope: 7.3 cm over 272 m•36-minute campus loop: 9.83 cm over 2.4 kmOutdoor accuracy is measured against a NovAtel RTK GNSS/INS reference.
If you’re working on robots that need reliable localization, from labs to farms, we’d love to connect.