SLAM
01
AVL SLAM System
Multi-sensor SLAM pipeline for an autonomous ground vehicle — fusing LiDAR, stereo cameras, and IMU through RTAB-Map on Jetson AGX Orin.
Year
2025 — present
Role
Research Assistant, Autonomous Vehicle Laboratory
Stack
Python · ROS 2 Humble · RTAB-Map · Jetson AGX Orin · Velodyne VLP-16 · ZED X · RealSense D455 · Xsens MTi-680G
Technical Highlights
- 01Sensor suite: Velodyne VLP-16, three ZED X stereo cameras, RealSense D455 RGB-D, and Xsens MTi-680G IMU running on Jetson AGX Orin / JetPack 6
- 02RTAB-Map configured with LiDAR-ICP odometry, RGB-D loop closure, and IMU fusion for robust pose estimation
- 03Multicamera loop closure — all cameras contribute equally to mapping via multicam node selection logic
- 04Flexible launch modes: 3-camera, 4-camera, and single-camera configurations with localization against saved map databases
- 05Resolved 13 major integration issues across realsense-ros 3.x → 4.x API migration, IMU topic remapping, and GPU parameter serialization conflicts