RNN for Perturbed Manipulability Optimization of Manipulators Based on a Distributed Scheme: A Game-Theoretic...

Area of Science:

• Robotics
• Control Systems
• Artificial Intelligence

Background:

• Redundant manipulators offer advantages but face challenges with singularities.
• Cooperative control of multiple manipulators is crucial for complex tasks.
• Singularity avoidance is fundamental for reliable motion planning and control.

Purpose of the Study:

• To propose a distributed scheme for improving the manipulability of redundant manipulators in a group.
• To address singularity avoidance during cooperative motion planning and control.
• To enhance the performance of multiple redundant manipulators in distributed networks.

Main Methods:

• Incorporating a manipulability index into cooperative control.
• Formulating the problem as a Nash equilibrium using game theory.
• Developing an anti-noise neural network to approximate the Nash equilibrium strategy.

Main Results:

• The proposed neural network model demonstrates superior global convergence and noise immunity.
• Simulations confirm the effectiveness of the neural network for real-time cooperative motion generation.
• The distributed scheme successfully guides manipulators to optimal spatial positions.

Conclusions:

• The developed neural network-based distributed scheme effectively improves cooperative motion control for redundant manipulators.
• The approach enhances manipulability and avoids singularities in perturbed distributed network environments.
• This method offers a robust solution for real-time cooperative motion generation in multi-manipulator systems.