LiDAR-Based Negative Obstacle Detection for Unmanned Ground Vehicles in Orchards

Area of Science:

• Robotics
• Agricultural Engineering
• Computer Vision

Background:

• Autonomous robots in orchards face safety risks from negative obstacles like ditches and potholes.
• Existing detection methods may struggle with the specific challenges of uneven orchard terrain.

Purpose of the Study:

• To develop and validate a novel method for detecting negative obstacles in orchard environments using LiDAR.
• To enhance the safety and reliability of autonomous robots operating in agricultural settings.

Main Methods:

• Utilizing a tilted Light Detection and Ranging (LiDAR) sensor (40° tilt) to improve ground point cloud data.
• Analyzing geometric features of laser point clouds, including rear wall height, density, and point spacing.
• Developing a mathematical model integrating multi-frame point cloud data for robust detection.

Main Results:

• Reduced blind spot from 3 m to 0.21 m and increased ground point cloud density.
• Achieved a 92.7% success rate in detecting negative obstacles in experimental trials.
• Established a maximum detection distance of approximately 8.0 meters for negative obstacles.

Conclusions:

• The proposed LiDAR tilt mounting method effectively detects negative obstacles in orchards.
• This technology significantly enhances robot safety and contributes to advancements in orchard automation.
• The method offers a practical solution for mitigating risks associated with ground-level hazards for agricultural robots.