GM Cruise Averaged 5,200 Miles Between AV Disengagements3
Fully autonomous vehicles are inevitable; however, they’re still likely decades away from mass production and adoption. Until then, cars with semi self-driving capabilities will fill the void, slowly making consumers comfortable with the idea of an electric chauffeur. Automakers and technology companies are all trying to be the first to introduce fully autonomous vehicles to the world, even if many prospects are limited in their capability.
One way to compare the progress of one company with that of another is to look at disengagements. In the AV industry, disengagements are instances where a human driver has to intervene and take control of the vehicle. Fewer disengagements equal bragging rights, and for GM Cruise, the autonomous vehicle subsidiary of General Motors, 2018 saw a huge improvement in the miles driver between disengagement in California.
Last year, GM Cruise recorded 447,621 miles driver with its autonomous fleet. Of those miles driven, the company recorded just 86 disengagements that required human intervention. This averages out to about 5,200 miles driven between disengagements, which is a considerable improvement over 2017’s 1,230 miles driven between disengagements. While GM Cruise did improve, it’s not the best out there. Waymo, a main competitor from Alphabet (Google), logged approximately 11,000 miles between disengagements, according to The Detroit News.
Over the last year, GM Cruise has focused on several of its testing locations, including San Francisco, Phoenix, and the Warren Technical Center. GM Cruise spokesman Milin Mehta said the company expanded operations equally across all locations. GM hopes to launch a ride-sharing service later this year with fully autonomous vehicles, and San Francisco will be the likely city. At the end of last month, the company published a short yet impressive video of a GM Cruise vehicle navigating tricky San Francisco streets. On several occasions, the autonomous vehicle was capable of crossing a double-yellow line to zip around traffic. It drove in a human-like fashion.
For 2019, GM Cruise expected to launch its ride-sharing service, and expects to double its resources, open a new office in the Seattle area, beginning its partnership with DoorDash, the meal delivery company. All that’s coming with General Motors earmarking $1 billion for the company in 2019—which is after it came in under budget in 2018.
All the while, rumors of a GM Cruise IPO have been catching fire.
- Sweepstakes Of The Month: Win a 2023 Corvette Z06 Convertible. Details here.
This only tells a small part of the story. I bet cruise could get much longer periods between disengagements if it’s highway only, aka controlled conditions. The question is where are the disengagements happening, what is tripping up the software etc. Also, what is the ‘feel’ of the driving with the autonomous software. A car can really only do 3 things, accelerate, brake or turn. Conservative software would brake whenever the software is unsure, so while technically not a disengagement, not exactly practical either. Autonomous cars will perform at the lowest common denominator, so if a certain scenario confuses the software (construction, double parked car, etc), then it won’t work. My informed forecast is first we gave geo fenced autonomous routes, with human controllers along the way (to deal with the lowest common denominator issues). Slowly that area is expanded, with no service in inclement weather. Full autonomous driving lanes on the highway are possible within a decade, but fully autonomous car you buy with no steering wheel, beyond 10 years.
These are good numbers that keep getting better.
The primary objective for major car manufacturers should be to become mobility companies owning huge AEV car fleets. These fleets have only to be sufficiently autonomous (level 4) to compete with taxis fleets, Uber and the likes. Waymo, GM Cruise and others are getting close. In 3 to 5 years these companies will have AEV car fleets with millions of cars, perhaps tens of millions of cars. In 5 to 8 years, these AEVs will displace many tens of millions of ICE cars.