A Secure Distributed Estimation Based on Affinity

This paper explores the issue of secure distributed estimation in the presence of attacks. Due to the presence of attacks, compromised nodes send erroneous information to their neighboring nodes, leading to a decline in network performance. Here, a distributed diffusion least-mean-square (DLMS) algorithm with affinity is proposed. The proposed algorithm uses affinity for secure distributed estimation. Specifically, we propose a distributed discrimination mechanism based on affinity to distinguish between normal nodes and malicious nodes. Simulations are conducted to demonstrate the superiority of the proposed DLMS-Affinity algorithm in comparison to other algorithms.