Exploring Autonomous Vehicle Safety: A Reliability Analysis Powered by JMP (2025-US-30MP-2560)

Artificial intelligence (AI) has become increasingly popular and widely adopted across various fields, with autonomous vehicles (AVs) emerging as one of its most visible and transformative applications. As AI technologies continue to advance, establishing the reliability of AVs is important for building public trust. Can we trust the safety of AVs? How does autonomy relate to safety? And what types of data can support such critical decisions?

In this talk, we focus on recurrent event data – specifically, collision events reported in the California Department of Motor Vehicles AV testing program – to assess AV safety performance. We demonstrate how to extract and analyze this publicly available data using JMP. Our analysis involves data exploration, potential cleaning, and application of reliability modeling techniques.

We also explore the relationship between disengagement events (discussed in a previous Discovery talk) and collision events to better understand how autonomy may impact safety. This study demonstrates how accessible data, when combined with JMP’s intuitive tools, can offer valuable insights into the reliability of AVs.