Leader-Following Formation of Heterogeneous Multi-Agent Systems with Time-Varying Topology: A Virtual Neighbor Framework
Date
2025
Authors
Xiao, Z.
Xu, H.
Tao, J.
Lu, R.
Shi, P.
Editors
Advisors
Journal Title
Journal ISSN
Volume Title
Type:
Journal article
Citation
IEEE Transactions on Automation Science and Engineering, 2025; 22:21341-21352
Statement of Responsibility
Zehui Xiao, Hongyi Xu, Jie Tao, Renquan Lu, Peng Shi
Conference Name
Abstract
This paper presents a set-membership control approach for formation tracking in heterogeneous leader-following multi-agent systems. In order to improve the adaptability of the control strategy to changes in topology, a virtual neighbor is constructed for each follower, which represents the combined influence of all its neighbors. With the help of virtual neighbors, the interaction relationships between agents can be hidden, thereby mitigating the impact of topology changes on controller design. Then, a novel time-varying topology with adaptive adjacent weights is proposed, such that the coupling strength between agents can be actively adjusted to improve the formation performance. Subsequently, the set-membership technique is utilized to establish the optimization problems to ensure that the one-step forward tracking errors can be constrained within the ellipsoid sets. Further, an online algorithm is developed to continuously obtain the optimal controllers by solving the optimization problems recursively. Finally, the effectiveness and practicality of the proposed approach are demonstrated via a numerical simulation and its application on a robotic platform with UAVs and UGVs. Note to Practitioners—Formation tracking in multi-agent systems, such as those involving UAVs and UGVs, often faces challenges brought by dynamic interaction topology and limited local information. To better adapt to the dynamics of the topology, the original interaction relationships of each agent will be concealed by introducing a virtual neighbor. Then, each agent only needs to consider the interactions with virtual neighbor, which simplifies the design of the controller and is beneficial for scaling up the system size. Another practical limitation in such scenarios is that agents often only have access to local information, making it unclear how to determine the best direction to move. To solve this problem, a time-varying topology with adaptive neighbor weights is proposed, which allows the agent to dynamically adjust its coupling strength according to the state information of its neighbors, so as to make more appropriate movement decisions in real time. Experimental results show that the proposed methods can adapt to continuous changes in topology, while effectively reducing formation errors. Moreover, considering the possible existence of obstacles in the actual environment, collision avoidance strategies will hopefully be incorporated to enrich the proposed virtual neighbor framework with time-varying topology.
School/Discipline
Dissertation Note
Provenance
Description
Access Status
Rights
© 2025 IEEE. All rights reserved, including rights for text and data mining, and training of artificial intelligence and similar technologies. Personal use is permitted, but republication/redistribution requires IEEE permission.