Game theory based vision impedance control for human-robot interaction.

Area of Science:

• Robotics
• Control Systems
• Human-Robot Interaction

Background:

• Accurate position tracking is crucial in human-robot interaction (HRI) systems.
• Multiple complex forces can negatively impact tracking precision.
• Existing control methods face challenges in managing these forces effectively.

Purpose of the Study:

• To propose a novel vision impedance control method for HRI systems.
• To address the influence of multiple forces on position tracking accuracy.
• To develop an optimal impedance controller using game theory.

Main Methods:

• Utilizing vision feedback and adaptive impedance law to estimate human-robot interaction forces.
• Converting position error into a constraint force to maintain output position limits.
• Applying a multi-party cooperative differential game algorithm to derive the optimal impedance controller.
• Proving convergence of position error using Lyapunov functions.

Main Results:

• The proposed method flexibly adjusts force proportions on the end-effector during HRI.
• Achieved mean squared error of end-effector position was 7.7% and 6.0% of sensorless force estimation and repetitive impedance learning methods, respectively.
• Demonstrated reduction in computational complexity compared to conventional vision methods.
• Validated improved tracking accuracy within constraints through simulations and experiments.

Conclusions:

• The developed vision impedance control method effectively manages multiple forces in HRI.
• The game theory-based approach enhances position tracking accuracy and reduces computational load.
• This method offers a flexible and efficient solution for complex HRI scenarios.