A visual servo tracking method for robotic arms based on adaptive neural network compensation
Patent Information
- Application Number
- CN202411462428.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-18
- Publication Date
- 2025-09-23
- Estimated Expiration
- 2044-10-18
AI Technical Summary
In traditional dynamic target visual servo tracking control, the image feature error is inaccurate due to the delay of the visual servo system and environmental noise, resulting in tracking accuracy. It is difficult to improve the control frequency of the visual servo system in the existing technology, and the tracking accuracy is low.
Adaptive neural network compensation is introduced. Real-time image feature error correction is performed in the robot arm visual servo system through the adaptive neural network compensator. The actual joint velocity of the robot arm and the image feature error are used for adaptive adjustment to improve the compensation accuracy.
Without increasing the control frequency of the visual servo system, the visual servo tracking accuracy of the robotic arm is significantly improved, the tracking error is reduced, and a more efficient tracking effect is achieved.
Smart Images

Figure CN119658673B_ABST
Abstract
Citation Information
Patent Citations
Mechanical arm target tracking method based on visual servo
CN112847334A
Visual servo finite time control method based on speed observer
CN118322205A