Label-Free Formation Control of Multi-Agent Systems with Bearing-Based Self-Organizing Map Neural Network

Hongyu Ji,Quan Yuan,Xiang Li
DOI: https://doi.org/10.1016/j.ifacol.2023.10.1497
2023-01-01
IFAC-PapersOnLine
Abstract:This paper proposes a label-free bearing-based formation control method of multi-agent systems based on a self-organizing map (SOM) neural network. The Follower Henneberg Construction is defined to construct the interaction network of the multi-agent system. The position of each agent in the target formation is not preset, but determined by the consensus-based auction algorithm (CBAA) in the competition stage of the SOM to obtain optimal trajectories of agents. The proposed method drives agents to form the target formation by controlling the bearings between each agent and its neighbors. The effectiveness and robustness of the method are verified by simulating the formation control process in 2-dimensional and 3-dimensional Euclidean space.
What problem does this paper attempt to address?