California DMV data shows Apple ranks dead last in autonomous vehicle miles between disengagements

“The Disengagement Reports for the period December 2017 to November 2018 for all entities operating under a license testing autonomous vehicles on public roads in California is/was out. As the DMV has taken off the reports after a short period (to be republished later this week), the following is what we got so far,” Mario Herger reports for The Last Driver License Holder. “These 28 companies operated over 321 vehicles (a couple of numbers, including Waymo and Zoox, are missing) for autonomous test drives in California. Many of the companies operate many more test cars in other states and countries.”

“In total all companies drove during that period 2,009,474 miles (3,215,159 kilometers) in autonomous mode and encountered 73,550 disengagements (the moment the system hands back control to a safety driver or the safety driver interferes),” Herger reports.

In “Miles per Disengagement,” Apple was dead last in the list with 1.1 miles. Of the total disengagements listed (73,550), Apple had, by far, the most, accounting for 69,510 of disengagements.

Read more in the full article here.

MacDailyNews Take: This is raw data, of course. It tells us nothing of the testing tolerances Apple has in place versus the others, for example.

[UPDATE: 3:44pm EDT: Added “miles between” to headline for clarity.]


        1. The post above was not posted by me, the real and registered KingMel. If you see a post that contains comments or opinions that differ from the positions that I have established over many years, then you can be assured that it is an imposter too weak-minded to generate a reputation for trustworthy and relevant MDN forum participation.

          These forum trolls are posing as legitimate registered users to sow discord and uncertainty in a forum already rife with political and social insanity.

  1. MDN: your headline is misleading. If Apple had ranked dead last in disengagements, it would mean that they had the LEAST number of disengagements. In fact, they had the MOST number of disengagements per mile driven. A better headline would have been: Apple has the highest number of autonomous vehicle disengagements. (the other way to look at this is Apple ranks dead last having the lowest distance traveled before a disengagement)

    1. I’ve been one to criticize MDN in the past for altering the headline of the original article, often to bring it more into line with their political bias. But this one is just plain bad construction. The original headline reads:

      Disengagement Reports 2018 – Preliminary Results

  2. This information is meaningless without knowing the details of how each company’s R&D effort works. Apple’s AI learning system may be quite different and with much stricter rules for disengagement.

    1. I mostly agree. The data is not sufficiently telling and you can only take so much from the comparisons without more context (severity/sensitivity of disengagement). That being said though, the message that Apple is not yet as successful as their peers is very probably a correct statement given that something has a need/reason to take over after only a mile of driving on average, this many years into development.

      I won’t read too much into it honestly. I’d also haircut the results from Google (1 every 11k miles) given that I understand that many of their miles driven are via preprogrammed routes. I’m not being critical of Google…but just trying to point out that we can’t easily draw strong conclusions from these numbers yet.

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.