Report issued on state of intelligent vehicle dependability and security
Intelligent road vehicles hold great promise for improving the efficiency and safety of our transportation systems. Such vehicles, however, can instead pose great risk if not deployed in a manner that is grounded in dependability and security.
A new report, the product of a five-year effort and co-authored by a group including two U-M EECS faculty, summarizes the activities of the Intelligent Vehicle Dependability and Security (IVDS) project, motivated to help industry move toward the development of intelligent vehicle control systems while measuring up to the standards of safety and security that have been established in other industries, such as aircraft.
The project was initiated in June 2019 by the International Federation for Information Processing (IFIP) Working Group 10.4 on Dependable Computing and Fault Tolerance, which Emeritus Prof. of EECS John Meyer helped to establish in 1980. The vision of the IVDS project’s members, according to Meyer, an author of the report, has been “the realization of highly dependable and secure operation of intelligent vehicles, verified and validated with respect to strict dependability, safety, and security requirements by rigorous state-of-the-art methods.”
The project was conducted through the engagement of stakeholders to increase awareness of dependability and security requirements, promote technical solutions, and provide expert help to governance and regulatory bodies in their rulemaking and oversight roles. The project concluded in early 2024.
Over the course of the project, representatives of stakeholder groups were brought together in two workshops to candidly discuss the current state of intelligent vehicle development and the challenges to realizing that vision. Additional engagements with the community included opinion pieces, white papers, academic journal publications, social media presence, and an international panel celebrating IFIP’s 60th Anniversary in October 2021.
Principal findings of the project point to significant shortfalls in the technologies, cost, governance, and societal aspects of achieving the end goal of safe and secure self-driving intelligent vehicles, generally referred to as Society of Automotive Engineers (SAE) Level 4 (full automation, subject to Operational Design Domain (ODD) constraints) or Level 5 (full automation).
“We don’t believe that widespread deployment of Level 4 or Level 5 vehicles on public roads is feasible in the near future,” said Carl Landwehr, a U-M Lecturer in Electrical and Computer Engineering and an author of the report. “Self-driving under highly restrictive ODD constraints – limitations designed to regulate automation based on environmental factors – appears to be a more feasible goal.”
All the project outputs, including the final report, workshop presentations and video recordings by leading researchers and practitioners in the field, have been organized into an easily accessible website ivds.dependability.org.