A Lightweight Localization Strategy for LiDAR-Guided Autonomous Robots with Artificial Landmarks.

Area of Science:

• Robotics
• Computer Vision
• Sensor Fusion

Background:

• Accurate robot localization is crucial for autonomous navigation.
• Traditional methods struggle with high-speed or dynamic environments.
• LiDAR-based odometry compensation is needed for enhanced precision.

Purpose of the Study:

• To develop a lightweight, real-time localization system (SORLA) using LiDAR and artificial reflectors.
• To compensate for laser odometry errors in high-speed and sharp-turning scenarios.
• To provide a robust and precise navigation solution for industrial applications.

Main Methods:

• Implemented a reflector extraction algorithm for precise localization of artificial landmarks from 2D LiDAR data.
• Developed a motion compensation technique to predict robot pose without odometer errors.
• Utilized a reflector matching optimization algorithm to ensure correct landmark identification during dynamic movements.

Main Results:

• Achieved an average localization error of 6.45 mm at 0.4 m/s and 9.87 mm at 4.2 m/s.
• Demonstrated robust performance with angular velocities up to 1.4 rad/s during sharp turns.
• Validated long-term stability over 72 hours and successful handling of reflector occlusion.

Conclusions:

• The SORLA system offers high precision and robustness for real-time robot localization.
• It is a feasible strategy for industry-level navigation, outperforming traditional SLAM in challenging conditions.
• The system presents a promising alternative for autonomous navigation systems requiring high accuracy.