GNSS/IMU/ODO/LiDAR-SLAM Integrated Navigation System Using IMU/ODO Pre-Integration

Area of Science:

• Robotics and Autonomous Systems
• Geomatics Engineering
• Navigation and Positioning

Background:

• Accurate and robust navigation is critical for autonomous systems, especially in environments with degraded Global Navigation Satellite System (GNSS) signals.
• Existing integrated navigation systems often struggle with sensor drift and environmental limitations, impacting performance in challenging scenarios like tunnels.

Purpose of the Study:

• To propose and evaluate a novel multi-sensor integrated navigation system fusing GNSS, Inertial Measurement Unit (IMU), Odometer (ODO), and LiDAR-SLAM.
• To enhance navigation accuracy and robustness by mitigating sensor drift and improving performance in GNSS-denied or feature-poor environments.

Main Methods:

• Developed a front-end dead reckoning system using IMU/ODO, incorporating odometer data to reduce IMU drift.
• Implemented a back-end graph optimization fusing GNSS, IMU/ODO pre-integration, and LiDAR-SLAM relative pose, utilizing a sliding window approach.
• Conducted land vehicle tests in open-sky and tunnel scenarios to validate system performance.

Main Results:

• The proposed system demonstrated significant reductions in navigation drift during simulated GNSS outages compared to conventional GNSS/INS/ODO integration (e.g., 62.8% North, 72.3% East, 52.1% Height position error reduction).
• Yaw error was reduced by 62.1%.
• In GNSS/IMU/LiDAR-SLAM integration, odometer assistance reduced vertical error by 72.3%. The system maintained short-term SLAM in tunnels and reduced forward positioning drift.

Conclusions:

• The multi-sensor fusion of GNSS, IMU, ODO, and LiDAR-SLAM effectively improves navigation accuracy and robustness.
• The integration of odometer data is crucial for mitigating drift, particularly in environments with limited GNSS availability and insufficient features for LiDAR-SLAM.
• The proposed system offers a reliable solution for autonomous navigation in challenging conditions where GNSS signals are weak and environmental features are sparse.