Prevention of false data injections in smart infrastructures

Abstract

Smart infrastructure is being developed on the basis of deep integration of cyberphysical systems and telecommunication networks to form a customer-oriented cyberspace. It acquires the ability to analyze the state of the entire system in real time to produce information and control processes in it. For splitting the cybernetic and ITC spaces, smart infrastructure is more vulnerable to security and safety threats than a static inter-computer network. This paper discusses an approach for detecting a specific safety threat, the false data injection (FDI), targeted at the smart infrastructures of such types as industrial Internet of Things (IIoT), smart buildings, e-hospitals, smart grids. This is a description of the proposed approach to identify the FDI attacks applying our new method that integrates a set of machine learning techniques. New approach has demonstrated 97% of accuracy at FDI prevention.