Privacy-Preserving Machine Learning Models for Traffic Forecasting

Details

• Creators:
  Mahmoud, Mahmoud N.
• Corporate Creators:
  North Carolina Agricultural and Technical State University
• Corporate Contributors:
  Center for Automated Vehicles Research with Multimodal Assured Navigation (CARMEN+) Tier-1 University Transportation Center (UTC) ; United States. Department of Transportation. University Transportation Centers (UTC) Program ; United States. Department of Transportation. Office of the Assistant Secretary for Research and Technology
• Subject/TRT Terms:
  Artificial Intelligence Autonomous Vehicles Data Files Functional Encryption Intelligent Transportation Systems Machine Learning Privacy Privacy-preserving Machine Learning Traffic Forecasting
• Resource Type:
  Tech Report
• Geographical Coverage:
  United States
• Edition:
  Final Report (June 2023 to Oct 2025)
• Corporate Publisher:
  Center for Automated Vehicles Research with Multimodal Assured Navigation (CARMEN+) Tier-1 University Transportation Center (UTC)
• Abstract:
  This report presents a comprehensive investigation into privacy-preserving traffic management, explainable artificial intelligence for autonomous systems, and cybersecurity in AV control. The research addresses critical challenges facing Intelligent Transportation Systems (ITS) and autonomous vehicles through four interconnected contributions. We introduce a secure and privacy-preserving traffic forecasting framework that combines Inner Product Functional Encryption (IPFE) with k-anonymity mechanisms to protect driver location data while enabling accurate traffic flow prediction through a hybrid deep learning architecture. We apply Concept Relevance Propagation (CRP), a bias-resistant explainable AI technique, to provide transparent concept-level explanations for traffic detection models in autonomous vehicles, enhancing trust and interpretability. We leverage CRP-generated explanations to automate dataset annotation for perception models, significantly reducing manual labeling effort while producing datasets that yield superior model performance. Finally, we develop an This report presents a comprehensive investigation into privacy-preserving traffic management, explainable artificial intelligence for autonomous systems, and cybersecurity in AV control. The research addresses critical challenges facing Intelligent Transportation Systems (ITS) and autonomous vehicles through four interconnected contributions. We introduce a secure and privacy-preserving traffic forecasting framework that combines Inner Product Functional Encryption (IPFE) with k-anonymity mechanisms to protect driver location data while enabling accurate traffic flow prediction through a hybrid deep learning architecture. We apply Concept Relevance Propagation (CRP), a bias-resistant explainable AI technique, to provide transparent concept-level explanations for traffic detection models in autonomous vehicles, enhancing trust and interpretability. We leverage CRP-generated explanations to automate dataset annotation for perception models, significantly reducing manual labeling effort while producing datasets that yield superior model performance. Finally, we develop an explainability-guided detection framework for trojan backdoor attacks in regression-based AV steering networks, achieving high detection rates for visible triggers and strong resilience against stealthy invisible variants. Together, these contributions establish new benchmarks for trustworthy AI in transportation, addressing fundamental challenges in data privacy, model transparency, and system security while demonstrating practical applicability for real-world deployment-guided detection framework for trojan backdoor attacks in regression-based AV steering networks, achieving high detection rates for visible triggers and strong resilience against stealthy invisible variants. Together, these contributions establish new benchmarks for trustworthy AI in transportation, addressing fundamental challenges in data privacy, model transparency, and system security while demonstrating practical applicability for real-world deployment.
• Format:
  PDF
• Funding:
  69A3552348327
• Collection(s):
  University Transportation Centers
• Main Document Checksum:
  urn:sha-512:62a3de7a5265e94003287b439b68f8e53d12a0cd0ca1389bddd46cc58d0c298920d48ee2e55942f313093625aa2ffcf050a806cfcda7c3f88988c51ab90b3eef
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/87923/dot_87923_DS1.pdf
• File Type:
  [PDF - 17.04 MB ]