Automatic Leader-Follower Persistent Formation Generation With Minimum Agent-Movement in Various Switching...

Area of Science:

• Robotics and Control Systems
• Multiagent Systems (MASs)
• Distributed Control

Background:

• Persistent formations are crucial for coordinated multiagent systems (MASs) but pose challenges in dynamic environments.
• Existing methods often struggle with maintaining formation integrity under switching topologies and minimizing agent displacement.

Purpose of the Study:

• To develop a novel generation strategy for relation-invariable persistent formations (RIPFs) in MASs.
• To enable efficient motion planning and control for achieving desired RIPFs with minimal agent movement.
• To address the complexities introduced by switching topologies in MAS formations.

Main Methods:

• An algorithm to determine the generability of persistent formations from rigid graphs.
• Introduction of state and transition matrices for managing RIPFs under switching topologies.
• A downward-tree combinatorial optimization algorithm for minimizing agent movement during formation changes.
• Design of a control law to transition between initial and desired RIPFs.

Main Results:

• A validated algorithm for generating RIPFs from rigid graphs and arbitrary initial configurations.
• Demonstrated effectiveness of state and transition matrices in handling dynamic topology changes.
• The downward-tree algorithm successfully minimized agent movement across different RIPFs.
• Simulation results confirm the efficacy of the proposed generation method, control law, and optimization algorithm.

Conclusions:

• The proposed framework effectively generates and controls relation-invariable persistent formations in multiagent systems.
• The methods presented allow for robust formation maintenance despite switching topologies and minimize energy expenditure.
• This work provides a significant advancement in the field of coordinated multiagent control and formation strategies.