2020/01/24 by Matthew Porter, Porter, Matthew, Sidhartha Dey +11
Engineering · #Autonomous Vehicle Technology and Safety #FOS: Electrical engineering #FOS: Mathematics #Optimization and Control (math.OC) #Smart Grid Security and Resilience #Systems and Control (eess.SY) #Vehicular Ad Hoc Networks (VANETs) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2001.09859
openalex publication_date 2020/01/24 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
Cyber-physical systems (CPS) such as autonomous vehicles rely on both\non-board sensors and external communications to estimate their state.\nUnfortunately, these communications render the system vulnerable to\ncyber-attacks. While many attack detection methods have begun to address these\nconcerns, they are limited to linear time-invariant (LTI) systems. Though LTI\nsystem models provide accurate approximations for CPS such as autonomous\nvehicles at constant speed and turning radii, they are inaccurate for more\ncomplex motions such as lane changes, turns, and changes in velocity. Since\nthese more complex motions are more suitably described by linear time-varying\n(LTV) system models rather than LTI models, Dynamic Watermarking, which adds a\nprivate excitation to the input signal to validate measurements, has recently\nbeen extended to LTV systems. However, this extension does not allow for LTV\nsystems that require several steps before the effect of a given control input\ncan be seen in the measurement signal. Additionally, there is no consideration\nfor the time-varying effects of auto-correlation. Furthermore, a proof of\nconcept was only provided using simulations of a simplified model.\n This paper relaxes the requirement for inputs to be visible in a single step\nand constructs an auto-correlation normalizing factor to remove the effects of\nauto-correlation. In addition, Dynamic Watermarking is applied to a\nhigh-fidelity vehicle model in carsim and a 1/10 scale autonomous rover to\nfurther reinforce the proof of concept for realistic systems. In each case, the\nvehicle follows a predefined path with time-varying velocity and turning radii.\nA replay attack, which replays previously recorded measurements, is shown to be\ndetectable using LTV Dynamic Watermarking in a quick and repeatable manner.\n