State of Self-Driving and Driver Assistance

In October, General Motors’s Cruise LLC unit will be allowed to test autonomous vehicles on public roads in San Francisco without a safety driver. they are the fifth company to get this permission. Inc.-backed Zoox, Alphabet [google] Waymo LLC, Nuro and AutoX Technologies are the other companies permitted to field test fully autonomous cars.

Cruise is permitted to operate five vehicles without a driver behind the wheel. They will be allowed to conduct both daytime and nighttime tests on certain public roads with speed limits no greater than 30 miles per hour and only in fair-weather conditions.

Before the end of 2020, Cruise will be sending cars out onto the streets of SF — without gasoline and without anyone at the wheel.

Consumer Reports Self Driving Test Gave Points for Making Sure Human is Alert to Disengage

Consumer Reports tested self-driving systems on track and on public roads. GM’s Cadillac scored 69 points out of a possible 100, while the Tesla scored 57. A Lincoln Corsair equipped with Ford Motor Co’s Co-Pilot 360 system, finished third with 52.

Consumer Reports gave a lot of points to GM for a driver-facing infrared camera to make sure the driver is still paying attention to the road and is ready to take over manual control when necessary. GM self-driving is not rating higher for actual self-driving but for monitoring a human.

Consumer Reports rates Tesla Autopilot better than GM Supercruise in terms of capabilities and performance and ease of use. Consumer Reports likes GM Supercruise preventing the driver from engaging it on roads where GM did not have detailed pre-scanned maps and where they other training wheel mechanisms.

GM is planning to charge $25 per month its Supercruise after a 3 year testing period. Tesla is charging $8000 for Autopilot.

California DMV’s 2019 autonomous vehicle report:
GM retrofitted Chevy Bolts drove 831,040 miles in 2019, up from 447,621 miles in 2018, 131,676 miles in 2017, and 9,776 miles in 2016.

Cruise recorded disengagements for the first half of 2019 separate from those recorded in the second half of 2019, reflecting what it says was a combination of upgrades and revamped testing procedures that drastically cut down on disengagements per mile. In the first half of 2019, Cruise reported 43 disengagements (or 7,635 miles per disengagement) in 328,285 total miles driven. In the second half of 2019, it logged 25 disengagements (or 20,110 miles per disengagement) in 502,755 miles. They had 68 total disengagements in 2019, versus 86 disengagements in 2018 (once every 5,205 miles) and 105 disengagements in 2017.

Waymo’s disengagement rates dropped from 0.09 per 1,000 self-driven miles (or one per 11,017 miles) to 0.076 per 1,000 self-driven miles (one per 13,219 miles).

Waymo had 153 cars and 268 drivers driving 1.45 million miles in California in 2019. They had 1.2 million miles in 2018, 352,000 miles in 2017, and 635,868 miles in 2016. Waymo had 1,500 monthly active riders in Phoenix, Arizona for their commercial taxi service, Waymo One. At the end of 2019, Waymo autonomous Chrysler Pacificas and Jaguar I-Pace electric SUVs have driven tens of billions of miles through computer simulations and 20 million miles on public roads in 25 cities.

Tesla conducts its testing via simulation, in laboratories, on test tracks, and on public roads in various locations around the world, and that it “shadow-tests” its cars’ autonomous capabilities by collecting anonymized data from over 400,000 customer-owned vehicles “during normal driving operations.”

Tesla had about 2 billion miles of self-driving data at the end of 2019 and now has over 4 billion miles of self driving data.

In 3rd quarter of 2020, Tesla registered one accident for every 4.59 million miles driven in which drivers had Autopilot engaged. Driving Tesla’s without Autopilot but with active safety features, there was one accident for every 2.42 million miles driven. For those driving without Autopilot and without our active safety features, there was one accident for every 1.79 million miles driven. By comparison, NHTSA’s most recent data shows that in the United States there is an automobile crash every 479,000 miles.

Actually using a Tesla autopilot, the increased safety for driving a Tesla with Autopilot engaged comes from removing 20% of your attention and effort from keeping track of maintaining speed and maintaining lane. This gives the option for someone to put more attention to other drivers on the road. This would allow a driver to better anticipate a developing situation and avoid an accident.

In Q4 2019, Tesla said it registered one accident for every 3.07 million miles driven in which drivers had Autopilot engaged.

SOURCES- California DMV, Tesla, Bloomberg, Cruise Medium
Written By Brian Wang,

1 thought on “State of Self-Driving and Driver Assistance”

  1. This is a good comparison. Would be keen to see a more comparable set of numbers – say miles per disengagements for cityurban driving across vendors. The problem with Tesla numbers is that most of the miles logged are on freeways.

Comments are closed.