Development of a simulation platform for safety impact analysis considering vehicle dynamics, sensor errors, and communication latencies: Assessing cooperative adaptive cruise control under cyber attack. (December 2018)

• Record Type: Journal Article
• Authors: Cui, Lian; Hu, Jia; Park, B. Brian; Bujanovic, Pavle
• Is Part Of: Transportation research. Volume 97(2018)
• Page Start: 1
• Page End: 22
• Publication Date: 2018-12
• Subjects: Connected and automated vehicle simulation -- Safety -- Cooperative adaptive cruise control -- Vehicle dynamics -- Equipment failure; Transportation -- Periodicals; Transportation -- Technological innovations -- Periodicals; 388.011
• DOI: 10.1016/j.trc.2018.10.005
• Languages: English
• ISSNs: 0968-090X

Abstract

Highlights: Universal CACC evaluation platform that is compatible with any CACC controller. Provide quantitative safety performance evaluation. Consistent benchmark for various CACC controllers. Capable of modeling the impact of vehicle dynamics, sensor errors, and communication latencies. Shed light on the state-of-the-art CACC system's cyber security. Abstract: While safety is one of the most critical contributions of Cooperative Adaptive Cruise Control (CACC), it is impractical to assess such impacts in a real world. Even with simulation, many factors including vehicle dynamics, sensor errors, automated vehicle control algorithms and crash severity need to be properly modeled. In this paper, a simulation platform is proposed which explicitly features: (i) vehicle dynamics; (ii) sensor errors and communication delays; (iii) compatibility with CACC controllers; (iv) state-of-the-art predecessor leader following (PLF) based cooperative adaptive cruise control (CACC) controller; and (v) ability to quantify crash severity and CACC stability. The proposed simulation platform evaluated the CACC performance under normal and cybersecurity attack scenarios using speed variation, headway ratio, and injury probability. The first two measures of effectiveness (MOEs) represent the stability of CACC platoon while the injury probability quantifies the severity of a crash. The proposed platform can evaluate the safety performance of CACC controllers of interest under various paroxysmal or extreme events. It is particularly useful when traditional empirical driver models are not applicable. Such situations include, but are not limited to, cyber-attacks, sensor failures, and heterogeneous traffic conditions. The proposed platform is validated against data collected from real field tests and tested under various cyber-attack scenarios.