Cyber-Attack Detection in Socio-Technical Transportation Systems Exploiting Redundancies Between Physical and Social Data

TL;DR

This paper presents a novel cyber-attack detection method for intelligent transportation systems that leverages redundancies between physical sensor data and social data from mobile devices, validated through simulations.

Contribution

It introduces a dual-filter scheme combining physical and social data for cyber-attack detection in ITSs, with stability analysis and simulation validation.

Findings

Effective detection of cyber-attacks demonstrated in simulations.

Utilizes social data to enhance detection robustness.

Mathematical stability of filters proven using Lyapunov theory.

Abstract

Cyber-physical-social connectivity is a key element in Intelligent Transportation Systems (ITSs) due to the ever-increasing interaction between human users and technological systems. Such connectivity translates the ITSs into dynamical systems of socio-technical nature. Exploiting this socio-technical feature to our advantage, we propose a cyber-attack detection scheme for ITSs that focuses on cyber-attacks on freeway traffic infrastructure. The proposed scheme combines two parallel macroscopic traffic model-based Partial Differential Equation (PDE) filters whose output residuals are compared to make decision on attack occurrences. One of the filters utilizes physical (vehicle/infrastructure) sensor data as feedback whereas the other utilizes social data from human users' mobile devices as feedback. The Social Data-based Filter is aided by a fake data isolator and a social signal processor that translates the social information into usable feedback signals. Mathematical convergence properties are analyzed for the filters using Lyapunov's stability theory. Lastly, we validate our proposed scheme by presenting simulation results.