Off-campus UMass Amherst users: To download campus access dissertations, please use the following link to log into our proxy server with your UMass Amherst user name and password.
Non-UMass Amherst users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
Dissertations that have an embargo placed on them will not be available to anyone until the embargo expires.
Author ORCID Identifier
https://orcid.org/0000-0002-9490-0119
AccessType
Campus-Only Access for One (1) Year
Document Type
dissertation
Degree Name
Doctor of Philosophy (PhD)
Degree Program
Electrical and Computer Engineering
Year Degree Awarded
2024
Month Degree Awarded
February
First Advisor
C.Mani Krishna
Second Advisor
Israel Koren
Third Advisor
Dennis L. Goeckel
Fourth Advisor
Michael Knodler Jr.
Subject Categories
Other Computer Engineering | Other Electrical and Computer Engineering
Abstract
In this thesis we propose new techniques to enhance the privacy of connected and autonomous vehicles (CAVs) using mix zones in Intelligent Transportation Systems (ITS). CAVs use safety messages to navigate in cooperation with neighboring vehicles. As safety messages contain vehicle identity, privacy is a major concern in ITS. Usage of pseudonym protects vehicle identity as long as the pseudonym is changed frequently and pseudonym exchange is untraceable. A common place to exchange pseudonyms is a mix-zone at a traffic-light intersection with multiple traffic flows. Within the specified physical boundaries of a mix-zone, vehicles follow a simple protocol of maintaining silence and changing pseudonyms. However, studies show that there is a high correlation between entry time and entry lane with the exit time and exit lane. Therefore, new techniques are needed to protect pseudonym exchange. We first present motivation for pseudonym exchange and provide a description of approaches to create confusion for an attacker. Then we introduce a new mix zone protocol named, Anonymity Enhancing Mix-zone Protocol (AEMP). We list the shortcomings of the existing work and present parameters which impact the mix-zone protocol’s performance to improve anonymity of CAVs in ITS in the presence of a strong adversary. We evaluate the improvement in anonymity with size of a mix-zone, varying traffic flows, traffic volume and physical characteristics of an intersection compared to a baseline protocol. The results suggest that AEMP improves CAV’s anonymity in a multi-lane intersection compared to the baseline protocol. We evaluate the tradeoff between vehicles gaining anonymity within mix-zones and the loss of throughput and accrued travel delays in two different road networks: Boston, MA and midtown NYC, NY. We found that an aggressive increase in the number of mix-zones impacts traffic efficiency metrics negatively. The optimal placement of mix-zones in a road network is critical when attempting to ensure the privacy and security of vehicular communication. Mix-zones are typically located at high-traffic intersections with diverse traffic flows, and their effectiveness depends on a variety of factors, including the density of vehicles and the network topology. In this study, we introduce a novel metric, called mixability, which quantifies the suitability of intersections as mix-zones. We apply genetic and annealing algorithms, which are popular heuristic optimization algorithms, to determine the optimal placement of mix-zones in the network, with the objective of maximizing the number of anonymous vehicles. Our study aims to provide insights into the design and operation of mix-zones in a way that can enhance the privacy and security of vehicular communication. The subsequent question we investigate is adapting the placement of mix zones to changes in traffic flow throughout the day. Adapting mix-zone placement to changing traffic flow presents a significant challenge, as the optimal placement depends on traffic density and flow patterns. We use past city traffic to extract traffic patterns using a clustering algorithm. We use two heuristics to place mix zones to fit the estimated traffic in each time slot. We assess the heuristics by considering the average disparity between the optimal solution and the heuristic solutions. We further modify the heuristics to adapt to changing traffic flow throughout the day. We compare our approach with static mix-zone placement and schemes with a) varying traffic observation times to estimate traffic and b) threshold values that trigger traffic changes based on traffic volume. The results of our study demonstrate the effectiveness of our approach, which addresses a critical research challenge and presents a promising solution to enhance privacy and security in urban transportation systems. Next, we evaluate the effectiveness of existing anonymizing protocols for vehicular networks and present a taxonomy of the protocols. We analyze the strengths and weaknesses of the protocols against three critical criteria: attacker’s strength, safety, and traffic efficiency. Among all the approaches, mix-zone protocols exhibit more advantages than drawbacks. To provide a more in-depth analysis, we examine the ability of various mix-zone protocols’ mechanisms to untangle spatial and temporal correlations of changes in vehicle pseudonyms. Additionally, we investigate different mix-zone placement algorithms and their impact on the overall performance of the protocols. The results of this study can guide future research in designing more effective and efficient anonymizing protocols for vehicular networks.
DOI
https://doi.org/10.7275/36459049
Recommended Citation
Ravi, Nirupama, "PRIVACY ENHANCING TECHNIQUES FOR CONNECTED AUTONOMOUS VEHICLES IN THE INTELLIGENT TRANSPORTATION SYSTEM" (2024). Doctoral Dissertations. 3083.
https://doi.org/10.7275/36459049
https://scholarworks.umass.edu/dissertations_2/3083
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.