Addressing Safety and Security Challenges in ML-based AV Software Stack – Remote Operation Support and Balancing Trade-offs

Center for Connected and Automated Transportation/Research/Research from the University of Michigan/Addressing Safety and Security Challenges in ML-based AV Software Stack – Remote Operation Support and Balancing Trade-offs

Addressing Safety and Security Challenges in ML-based AV Software Stack – Remote Operation Support and Balancing Trade-offs

In Progress
Headshot of Z. Morley Mao. The link directs to their profile page.
Z. Morley Mao
Headshot of Yiheng Feng. The link directs to their profile page.
Yiheng Feng
The University of Michigan Logo. The link directs to the funded research led by this institution.
Purdue University Logo. The link directs to the funded research led by this institution.

Principal Investigator(s):

Z. Morley Mao, Professor of Electrical Engineering and Computer Science – The University of Michigan
Yiheng Feng, Assistant Professor of Civil Engineering – Purdue University

Project Abstract:
In this work, we propose a framework that enables safe and secure human remote operation when AV systems require support due to the inherent limitations of ML models used. We develop an approach that can effectively detect and mitigate potentially malicious remote human operators, and satisfy the real-time requirements of remote operation despite possibly variable network conditions impacting the communication channel between the AV system and the remote operator. Our solution will be demonstrated on the Mcity 2.0 testbed as a means to validate the proposed design in realistic settings.

Institution(s): University of Michigan — Ann Arbor
Purdue University

Award Year: 2024

Research Focus: Safety, Cybersecurity, Mobility

Project Form(s):

Project Information Form
Data Management Plan