Method for detecting unauthorized nodes

Abstract

This paper proposes an anomaly detection method based on a combination of Bayesian models for estimating the probability of node activity with stochastic filtering methods that provide adaptive real-time update of estimates. This approach allows implementing an effective security control system capable of detecting potentially dangerous nodes even in the absence of additional cryptographic mechanisms or centralized monitoring.