Showing posts from November, 2018

FiveAI Report on Autonomous Vehicle Safety Certification

FiveAI has published an autonomous vehicle safety approach that includes independent verification, transparency, and data sharing. (I provided inputs to the @_FiveAI  authors.) Here is a pointer to the summary on Medium It's worth jumping through the registration hoop to read the full version.

Webinar on Robustness Testing of Perception

Zachary Pezzementi and Trenton Tabor have done some great work on perception systems in general, and how image degradation affects things.  I'd previously posted information about their paper, but now there is a webinar available here:     Webinar home page with details & links: This includes pointers to slides, a recorded webinar, the paper, and papers. My robustness testing team at NREC worked with them on the perception stress testing parts, so here are quick links to the parts covering that part: Paper: robustness testing for safe perception Webinar slides 40-50: Video time 44:43 - 54:43:

Potential Autonomous Vehicle Safety Improvement: Less Hype, More Data (OESA 2018)

I enjoyed being on a panel at the Annual OESA Suppliers Conference today. My intro talk covered setting reasonable expectations about how much safety benefit autonomous vehicles can provide in the near-term to mid-term. Spoiler: when you hear that 94% of all road fatalities are caused by bad and impaired drivers, that number doesn't mean what you think it means! Autonomous Vehicle Safety Improvement: Less Hype, More Data from Philip Koopman

Uber ATG Safety Report

Summary: Uber's reports indicate that they are taking improving their safety culture seriously. Their new approach to public road testing seems reasonable in light of current practices. Whether they can achieve an appropriate level of system safety and software quality for the final production vehicles remains an open question -- just as it does for the self-driving car industry in general. Uber ATG has released a set of materials regarding their in-development self-driving car technology and testing, including a NHTSA-style safety report, as well as reports of a safety review in light of the tragic death in Tempe AZ earlier this year. (See: ) Generally I have refrained from critical analysis of other company safety reports because this is still something everyone is sorting out. Anyone putting out a safety report is automatically in the top 10% for transparency (because about 90% of the companies haven't even released one yet).