A software Tool for Securing Deep Learning against Adversarial Attacks for CAVs

The scope of this project is to realize a technology transfer of previously developed adversarial attack resilient perception algorithm for autonomous navigation in connected and automated vehicles. The project will attain its objectives by developing a graphical user interface (GUI) designed for the automatic generation of a robust perception algorithm. This involves utilizing a baseline perception neural network and offering options for training and evaluation using various publicly available image datasets. Furthermore, the initiative includes training students on the utilization of the developed GUI, aiming to cultivate a generation of proficient individuals poised to contribute to the advancement of the United States Department of Transportation (USDOT) and society as a whole. Practical applications and understanding of the proposed technology will be demonstrated through workshops, employing an autonomous F1/10 car testbed that emulates a Connected and Autonomous Vehicle (CAV) perception module. The technology’s viability will be showcased by implementing it on an autonomous vehicle within the AVL DrivingCube SCENIUS at CU-ICAR, highlighting its real-world applicability and potential impact on autonomous vehicle technology.

Language

• English

Project

• Status: Active
• Funding: $186260
• Contract Numbers:

  69A3551747117

• Sponsor Organizations:

  Office of the Assistant Secretary for Research and Technology

  University Transportation Centers Program
  Department of Transportation
  Washington, DC  United States  20590

  Benedict College

  1600 Harden Street
  Columbia, South Carolina  United States  29204

  Clemson University

  216 Lowry Hall
  Clemson, SC, SC  United States  29634
• Managing Organizations:

  Center for Connected Multimodal Mobility

  Clemson University
  Clemson, SC  United States  29634
• Performing Organizations:

  Clemson University

  216 Lowry Hall
  Clemson, SC, SC  United States  29634

  Benedict College

  1600 Harden Street
  Columbia, South Carolina  United States  29204
• Principal Investigators:

  Pisu, Pierluigi

• Start Date: 20231001
• Expected Completion Date: 20240930
• Actual Completion Date: 0
• USDOT Program: University Transportation Centers

Subject/Index Terms

• TRT Terms: Autonomous vehicles; Computer security; Connected vehicles; Education and training methods; Graphical user interfaces; Machine learning; Software
• Subject Areas: Data and Information Technology; Education and Training; Highways; Security and Emergencies; Vehicles and Equipment;

Filing Info

• Accession Number: 01907175
• Record Type: Research project
• Source Agency: Center for Connected Multimodal Mobility
• Contract Numbers: 69A3551747117
• Files: UTC, RIP
• Created Date: Feb 6 2024 8:12PM