Between Dec. 23, 2024 and Jan. 3, 2025, eCommons staff will not be available to answer email and will not be able to provide DOIs until after Jan. 6. If you need a DOI for a dataset during this period, consider Dryad or OpenICPSR. If you need support submitting material before the winter break, please contact us by Thursday, Dec. 19 at noon. Thank you!
Library

eCommons

 

Improving Security and User Privacy in Learning-Based Traffic Signal Controllers (TSC)

Simple item page
dc.contributor.authorHaydari, Ammar
dc.contributor.authorZhang, Michael Z.
dc.contributor.authorChuah, Chen-Nee
dc.date.accessioned2022-03-14T17:47:48Z
dc.date.available2022-03-14T17:47:48Z
dc.date.issued2022-02-28
dc.descriptionFinal Reporten_US
dc.description.abstractThe 21st century of transportation systems leverages intelligent learning agents and data-centric approaches to analyze information gathered with sensing (both vehicles and roadsides) or shared by users to improve transportation efficiency and safety. Numerous machine learning (ML) models have been incorporated to make control decisions (e.g., traffic light control schedules) based on mining mobility data sets and real-time input from vehicles via vehicle-to-vehicle and vehicle-to-infrastructure communications. However, in such situations, where ML models are used for automation by leveraging external inputs, the associated security and privacy issues start to surface. This project aims to study the security of ML systems and data privacy associated with learning-based traffic signal controllers (TSCs). Preliminary work has demonstrated that deep reinforcement learning (DRL) based TSCs are vulnerable to both white-box and black-box cyber-attacks. Research goals include 1) quantifying the impact of such security vulnerabilities on the safety and efficiency of the TSC operation, and 2) developing effective detection and mitigation mechanisms for such attacks. In learning based TSCs, vehicles share their messages with the DRL agents at TSCs, which will then analyze the data and take action. Sharing vehicular mobility data with a network of TSCs may cause privacy leakage. To address this problem, differential privacy techniques will be applied to the mobility datasets to protect user privacy while preserving the effectiveness of the prediction outcomes of traffic-actuated or learning-based TSC algorithms. Approaches will be evaluated in vehicular simulators using real mobility data from San Francisco and other cities in California. By accomplishing these goals, learning-based transportation systems will be more secure and reliable for real-time implementations.en_US
dc.description.sponsorshipU.S. Department of Transportation 69A3551747119en_US
dc.identifier.urihttps://hdl.handle.net/1813/111129
dc.language.isoen_USen_US
dc.rightsAttribution 4.0 International*
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/*
dc.subjectDeep reinforcement learningen_US
dc.subjecttraffic signal controlen_US
dc.subjectadversarial attacken_US
dc.subjectsecurity defenseen_US
dc.subjectstatistical anomaly detectionen_US
dc.titleImproving Security and User Privacy in Learning-Based Traffic Signal Controllers (TSC)en_US
dc.typereporten_US
schema.accessibilityFeaturereadingOrderen_US
schema.accessibilityFeaturestructuralNavigationen_US
schema.accessibilityFeaturetaggedPDFen_US
schema.accessibilityHazardunknownen_US

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
UCD_YR5_CHUAH_FINAL_IMPROVING SECURITY.pdf
Size:
361.82 KB
Format:
Adobe Portable Document Format
Description:
Download

Collections

CTECH Final Reports