Marked-LIEO: Visual Marker-Aided LiDAR/IMU/Encoder Integrated Odometry.

Area of Science:

• Robotics
• Computer Vision
• Sensor Fusion

Background:

• Accurate pose estimation is crucial for mobile robot navigation in complex indoor environments.
• Existing methods struggle with challenges like GNSS signal loss, wheel slippage, and LiDAR degradation in long corridors.

Purpose of the Study:

• To propose Marked-LIEO, a novel visual marker-aided LiDAR/IMU/encoder integrated odometry system.
• To achieve robust and accurate pose estimation for mobile robots in indoor long corridor environments.

Main Methods:

• A two-stage approach involving pre-integration of encoder and IMU data, and joint optimization of LiDAR and low-frequency visual marker odometry.
• Adaptive algorithm to adjust optimization weights based on yaw angle and LiDAR degradation distance.
• Multi-sensor fusion through joint optimization of encoder, IMU, LiDAR, and camera measurements.

Main Results:

• Marked-LIEO successfully achieves accurate pose estimation in indoor corridor environments.
• The system demonstrates robustness against wheel slipping and LiDAR degradation.
• Experimental validation in both Gazebo simulation and real-world environments confirms the method's effectiveness.

Conclusions:

• Marked-LIEO provides a reliable solution for mobile robot localization in challenging indoor corridor settings.
• The integration of visual markers and adaptive algorithms enhances pose estimation accuracy and stability.
• This approach addresses key limitations of traditional odometry systems in GNSS-denied and visually degraded environments.