Research on autonomous navigation system of greenhouse electric crawler tractor based on LiDAR

Area of Science:

• Agricultural Engineering
• Robotics
• Autonomous Systems

Background:

• Intelligent greenhouses require autonomous navigation for electric crawler tractors.
• Enclosed, narrow spaces and uneven terrain in greenhouses pose navigation challenges.
• Existing systems lack the intelligence for complex greenhouse environments.

Purpose of the Study:

• To develop an intelligent navigation system for electric crawler tractors in greenhouse environments.
• To improve the autonomy and precision of tractors in enclosed agricultural settings.
• To address challenges of narrow spaces, uneven ground, and obstacles.

Main Methods:

• A navigation system integrating LiDAR, Inertial Measurement Unit (IMU), and wheel speed sensors was developed.
• Gmapping algorithm for environment mapping and adaptive Monte Carlo localization for positioning were employed.
• A* algorithm for global path planning and Dynamic Window Approach (DWA) for local path planning and obstacle avoidance were utilized.

Main Results:

• The A* algorithm demonstrated superior global path planning, reducing inflection points by up to 87.50% and runtime by 81.49% compared to other algorithms.
• Map accuracy achieved a maximum lateral deviation of 6 cm and longitudinal deviation of 16 cm.
• Navigation accuracy showed maximum lateral deviation < 13 cm, average < 7 cm, and standard deviation < 8 cm; maximum heading deviation < 14°, average < 7°, and standard deviation < 8°.

Conclusions:

• The developed LiDAR-based autonomous navigation system effectively meets the accuracy requirements for electric crawler tractors in greenhouse environments.
• The combination of A* and DWA algorithms ensures optimal global paths and efficient local obstacle avoidance.
• The system enhances the intelligence and operational capabilities of tractors in precision agriculture.