一种基于安全强化学习的机械臂视觉伺服控制方法及系统
By introducing a control obstacle function (CBF) into the visual servo control of a robotic arm, the problems of feature point loss and low training efficiency in traditional methods are solved, and high-precision and high-safety visual servo control is achieved in multi-obstacle environments.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA UNIV OF MINING & TECH
- Filing Date
- 2026-04-27
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional visual servoing methods for robotic arms suffer from problems such as nonlinear mapping leading to system failure, feature point loss, low training efficiency, and insufficient security in environments with multiple obstacles and complex environments. Existing deep reinforcement learning methods have low sample efficiency and lack explicit security guarantees.
By introducing the Control Obstacle Function (CBF) into the entire reinforcement learning process, safe gradients and safe filters are constructed through the CBF function to ensure that feature points are within the field of view, thereby improving data collection and policy training, and correcting actions in real time to achieve highly safe obstacle avoidance.
It improves data acquisition efficiency, enhances learning stability and deployment security, achieves high-precision and high-security visual servo control, and reduces development and debugging costs.
Smart Images
Figure CN122077670B_ABST