弹性分布式参数估计算法研究进展

Progress in the study of resilient distributed parameter estimation algorithms

摘要:

基于自适应网络的分布式参数估计近年来受到了日益广泛的关注。现有的分布式参数估计算法尽管在无攻击的安全网络中表现良好，但在遭受如虚假数据注入（false data injection，FDI）攻击的对抗网络中，由攻击者注入的虚假数据（也称恶意数据）会随着节点间的通信和协作在网络中扩散，导致算法估计性能的恶化。若算法不能从攻击中快速恢复估计性能（即算法对攻击不具有弹性），则可能导致严重的后果。为此，简要介绍了弹性分布式参数估计算法所解决的基本问题及基本算法原理；从FDI攻击检测和弹性参数估计策略2个方面，系统地总结了近年来弹性分布式参数估计算法的研究进展，并分析了其在遭受FDI攻击的对抗网络中的性能；最后，探讨了现有弹性分布式参数估计算法的发展趋势和未来潜在的研究方向。

Abstract:

Distributed parameter estimation based on adaptive networks has received increasing attention in recent years. Although existing distributed parameter estimation algorithms perform well in secure networks without attacks, in adversarial networks subjected to attacks such as false data injection (FDI), the false data (also known as malicious data) injected by attackers will spread throughout the network through node communication and collaboration, leading to a deterioration of the algorithm’s estimation performance. If the algorithm cannot quickly recover its estimation performance from the attack (i.e., the algorithm is not resilient to the attack), it may lead to serious consequences. To this end, this paper first briefly introduced the basic problems and principles of resilient distributed parameter estimation algorithms; then, it systematically summarized the research progress of resilient distributed parameter estimation algorithms in recent years from two aspects: FDI attack detection and elastic parameter estimation strategies, and analyzed their performance in adversarial networks subjected to FDI attacks; finally, it discussed the development trend and potential future research directions of the existing resilient distributed parameter estimation algorithms.