Distributed and Time-Delayed k-Winner-Take-All Network for Competitive Coordination of Multiple Robots

Area of Science:

• Robotics
• Control Systems
• Network Theory

Background:

• Multirobot systems require efficient task assignment strategies.
• Existing coordination methods often face communication and computational limitations.
• Time delays are a common challenge in distributed network systems.

Purpose of the Study:

• To develop and analyze a distributed and time-delayed k-winner-take-all (DT-kWTA) network.
• To enable competitively coordinated task assignment in multirobot systems.
• To address communication load, computational burden, and time delays in network design.

Main Methods:

• Established a network based on the k-winner-take-all (kWTA) competitive algorithm.
• Implemented a distributed control strategy to optimize network performance.
• Incorporated time-delay analysis into the network design.

Main Results:

• The DT-kWTA network was successfully combined with time delay for the first time.
• Theoretical derivation of the maximum allowable delay was performed.
• Convergence and robustness of the network were mathematically proven.
• The network was applied to a multirobot system for task completion.

Conclusions:

• The DT-kWTA network offers improved capabilities for realistic applications compared to previous methods.
• The proposed network provides a robust and convergent solution for multirobot task assignment.
• This work advances the field of distributed control for coordinated robotic systems.