摘要
The consensus mechanism in multi-agent networks has attracted considerable attention in both control and computer science.However,current advancements in consensusbased control theory lack a general framework to optimize the communication complexity required to reach consensus.This gap highlights the necessity of robust analytical frameworks to advance the field.Our proposed method,termed hierarchical random networks,decomposes the entire network into multiple random sub-swarms and constructs a hierarchical structure among these sub-swarms.First,we establish a simplified condition to ensure the connectivity of hierarchical random networks.Further,we prove that the expected number of network connections in hierarchical random networks can be reduced to its lower bound as the size of sub-swarms approaches the square root of the total number of agents.At the end of the paper,we validate the effectiveness of the proposed network topology through simulation case studies on maneuvering target tracking.The results demonstrate that combining hierarchical random networks with consensus-based filters can achieve maneuvering target tracking while reducing communication complexity.
基金
support by the Aeronautical Science Foundation of China(20220001057001,20240001057002).
