Secure Distributed Estimation Over Multitask Networks Against Multiple Attacks

In recent years, distributed multitask estimation on networks, where different sensor nodes estimate different parameters, has received increasing attention. This article mainly focuses on the challenge of distributed security estimation in the context of multiple attacks on multitask networks. First, we propose a distributed multitask secure estimation algorithm based on the local outlier factor and intertask correlation in a multitask network. Additionally, a novel distributed time-varying fusion strategy that equalizes the weight assigned to each node depending on the data density is proposed to mitigate further the effect of multiple attacks on the algorithm's performance. Finally, the performance of the proposed algorithm is evaluated in terms of mean and mean square error, as well as simulated experiments against multiple attacks. Theoretical and numerical studies indicate that the proposed approach can efficiently repel multiple attacks.