Waymo deploys with no human safety driver oversight

Topic: 

In a major milestone for robocars, Waymo has announced they will deploy in Phoenix with no human safety drivers behind the wheel. Until now, almost all robocars out there have only gone out on public streets with a trained human driver behind the wheel, ready to take over at any sign of trouble. Waymo and a few others have done short demonstrations with no safety driver, but now an actual pilot, providing service to beta-testing members of the public, will operate without human capability for intervention.

This is a big deal, and indicates Waymo's internal testing is showing a very strong safety record. The last time they published numbers, they had gone 83,000 miles between "necessary interventions." While in safety driver training, we are told to intervene at any sign of a problem, these interventions are tested in simulation to find out what would have happened if there had been no intervention. If the car would have done the right thing, it's not a required intervention.

Waymo must have built their number up a great deal from there. People have an accident that is reported to insurance about ever 250,000 miles, and to police every 500,000 miles. Injury accidents happen every 1.2M miles, and fatalities every 80M miles. In Waymo's testing, where they got hit a lot by other drivers, they discovered that there are "dings" about every 100,000 miles that don't get reported to police or insurance.

People have argued about how good you have to be to put a robocar on the road. You need to be better than all those numbers. I will guess that Waymo has gotten the "ding" number up above 500,000 miles -- which is close to a full human lifetime of driving. Since they have only driven 3.5M miles they can't make real-world estimates of the frequency of injuries and certainly not of fatalities, but they can make predictions. And their numbers have convinced them, and the Alphabet management, that it's time to deploy.

Congratulations to all the team.

It should be noted that Waymo is keeping a human supervisor in the back seat. Some feel this seriously reduces the accomplishment, but this supervisor can only do simple things like tell the vehicle to pull off the road. They can't do tactical takeovers the way drivers behind the wheel could. Waymo has tons of money, so does not need to remove the human for cost reasons during this pilot, so they don't want to add any danger they can avoid. It may not be too long until these rear seat supervisors are gone (though remote control center operators will probably be on-call for decades.)

They did this not just with real world testing, but building a sophisticated simulator to test zillions of different situations, and a real world test track where they could test 20,000 different scenarios. And for this pilot they are putting it out on the calm and easy streets of Phoenix, probably one of the easiest places to drive in the world. Together, that gives the confidence to put "civilians" in the cars with no human to catch an error. Nothing will be perfect, but this vehicle should outperform a human driver. The open question will be how the courts treat that when the first problem actually does happen. Their test record suggests that may be a while; let us hope it is.

Where do we go from here?

This pilot should give pause to those who have said that robocars are a decade or more away, but it also doesn't mean they are full here today. Phoenix was chosen because it's a much easier target than some places. Nice, wide streets in a regular grid. Flat terrain. Long blocks. Easy weather with no snow and little rain. Lower numbers of pedestrians and cyclists. Driving there does not let you drive the next day in Boston.

But neither does it mean it takes you decades to go from Phoenix to Boston, or even to Delhi. As Waymo proves things out in this pilot, first they will prove the safety and other technical issues. Then they will start proving out business models. Once they do that, prepare for a land rush as they leap to other cities to stake the first claim and the first-mover advantage (if there is one, of course.) And expect others to do the same, but later than Waymo, because as this demonstrates, Waymo is seriously far ahead of the other players. It took Waymo 8 years to get to this, with lots of money and probably the best team out there. But it's always faster to do something the 2nd time. Soon another pilot from another company will arise, and when it proves itself, the land rush will really begin.

Comments

Amazing progress by Waymo...
You mention that Waymo's intervention rate is once every 83,000 miles. Where was this published? The last numbers I saw were about 5,000 miles...

This was a figure (at the time) of the farthest they have gone without an intervention necessary to prevent a "contact" (ie. hitting something.) They intervene for other reasons more often. They have clearly gotten a lot better since then. 83,000 miles is about 8 years of driving for the average person, but less than 2 years for the average taxi.

It is going to be fascinating to see whether Google can leverage the huge lead they have. The one piece of the puzzle that seems to be missing at the moment is who are they going to partner with to build the cars, if indeed they do.
In my mind they should be looking for partners in a number of areas.
Firstly the ride hailing/sharing partner, they seem to have settled on Lyft, though GM are also partnering with Lyft an appear to be one of the more credible competitors.
Secondly manufacturing the vehicles. I would be looking further afield than Chrysler to a global manufacturer such as Toyota, so that the vehicles can be rolled out to new markets quickly. GM could be an even stronger proposition as they are already some way down the path of making a Bolt explicitly for self driving. If GM and Google got together, and ditched Cruise, then you would think they could get a sizable jump on the market.
I am assuming of course that Google want to partner up in order to get the complete package to market earlier, rather than just sell their software to whoever sets up in this space and hope that they keep their lead.
It will also be interesting to see when self driving cars are manufactured for private sale, and by whom.

Google doesn't need a ride share partner. They are one company that could crush that business if they decided to enter. Apple's the other one. They don't need a manufacturing partner. All they need is a supplier. Like Foxconn is to Apple. There are lots of suppliers. The car companies on the other hand don't want to be just a supplier like Foxconn, they want to be a partner, so they may offer a lot to get a partnership instead of a purchase order. Chrysler would love to be that partner but they are not as yet.

Today at lunch we saw one of these: https://www.macrumors.com/2017/08/25/apple-new-autonomous-driving-suvs/ (only in black) An interesting approach would be to build the sensor/compute package as a unit that could be bolted on to many different vehicles. This would be ideal for low volume deployments that wouldn't require deep integration into the vehicle design and manufacturing process. And for robotaxis, would allow for rapid and frequent upgrades while still using the same base vehicle.

Generally, nobody is making a bolt-on. They want to extensively test each configuration, which means a new vehicle, custom modified. There is no virtue to being able to support a lot of different standard cars. (There is virtue in supporting different types of cars -- single person, sedan, SUV) but each would need to be custom designed and tested.

This seems to contradict the earlier statement that Google don't need a partner, they need a supplier. I would have thought that deep integration would be a whole lot easier if the car manufacturer was actively involved in developing the car specifically as an autonomous vehicle.
Certainly a city bound robotaxi is likely to be small light, low performance, and electric of which there are relatively few in the market place today.
They wouldn't want GM/Cruise to come out with inferior but adequate self driving software, but with a much more suitable (and cheaper) car and gain market share that way.

And you're rich like Google, you just pay a company to make a car to your specification. They are your supplier more than your partner. Most car companies do contract manufacturing like this.

If they are just looking for a supplier instead of an OEM partner why don’t they contract with Magna?

This is just amazing that Google is able to do this. Can't wait until they spread across the US and I can use this service.

Add new comment