Bioinspired Simultaneous Learning and Motion-Force Hybrid Control for Robotic Manipulators Under Multiple...

Area of Science:

• Robotics
• Control Theory
• Bio-inspired Systems

Background:

• Biological systems exhibit adaptive motion coordination.
• Robotic manipulators require precise and compliant control for dexterous interaction.
• Physically constrained environments pose challenges for robotic control.

Purpose of the Study:

• To develop a bioinspired control strategy for precise and compliant motion-force coordination in robotic manipulators.
• To enhance robotic robustness and precision in constrained environments.
• To enable embodied intelligence and dexterous interaction.

Main Methods:

• Proposed a learning-based motion-force hybrid control (LMFC) framework.
• Formulated motion-force coordination as a time-varying quadratic programming (TVQP) problem.
• Integrated an RNN-based controller for adaptive learning and online parameter estimation.

Main Results:

• The LMFC framework effectively regulates motion and interaction forces under incomplete kinematic information.
• Practical constraints (joint limits, orientation, obstacles) were incorporated at the acceleration level.
• The RNN controller mitigated joint drift and estimated uncertain kinematic parameters online.

Conclusions:

• The proposed LMFC framework demonstrates effectiveness and practicality for adaptive and compliant robotic control.
• The approach enhances robotic capabilities in constraint-rich environments.
• Bio-inspired strategies offer significant potential for advanced robotic applications.