基于深度强化学习的表型巡检机器人垄行跟踪导航方法及系统
By using deep reinforcement learning, a perceptual information processing and motion decision-making module was constructed, which solved the problems of poor navigation flexibility and significant light impact on agricultural robots. This enabled autonomous and flexible field navigation, reduced the workload of data collection, and adapted to the navigation needs of different crop fields.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HARBIN INST OF TECH AT WEIHAI
- Filing Date
- 2025-05-16
- Publication Date
- 2026-07-17
AI Technical Summary
Existing navigation methods for agricultural robots require extensive prior data collection, lack flexibility, are greatly affected by lighting conditions, and have poor usability.
A phenotypic tracking and navigation method for inspection robots based on deep reinforcement learning is adopted. By constructing a perception information processing module, a state dimensionality upgrading module, and a motion decision module, feature information is extracted using convolutional neural networks and autoencoder mechanisms. Lateral offset, yaw angle offset, sway change rate, and forward reward function are designed for training to achieve autonomous navigation.
It achieves flexible and autonomous navigation, eliminating the need for preset fixed paths, reducing the workload of preliminary data collection, minimizing the impact of sunlight, providing 24-hour navigation capability, adapting to the navigation needs of different crop fields, having a small number of network parameters, and simplifying program deployment.
Smart Images
Figure CN120563557B_ABST