Secure Distributed Estimation Based on Density Peak in Multi-Task Networks
Honghua Yi, Xiaoping Ren · 2024
Distributed multi-task parameter estimation has attracted considerable attention in recent years, finding applications across diverse domains. However, in open external environments, wireless sensor networks (WSNs) are particularly vulnerable to malicious attacks, which can compromise sensor data and degrade overall network performance. To achieve secure estimation, we propose a multi-task diffusion least mean square with density peak (MDLMS-DP) algorithm. This algorithm utilizes density peak nodes to establish adaptive thresholds, effectively identifying compromised nodes. Furthermore, we introduce an innovative data fusion strategy to improve estimation accuracy and robustness. Simulation results validate that the proposed algorithm maintains robust performance even in adversarial networks.