CruiseDriverlessCar

Vjeran Pavic

Mobile


Neither Google nor the big car manufacturers will likely be the first to put the keys to self-driving car technology in consumers’ hands. That honor will go to an eight-person startup called Cruise. Preorders start on Monday for installation in early 2015.

“A third of the American workforce spends more than an hour a day commuting. It’s boring and dangerous. Now that we have the technology, it’s almost our responsibility to do something with it,” Cruise CEO Kyle Vogt told Re/code.

Cruise makes an aftermarket kit for newer Audi cars that allows them to drive on autopilot on highways in the San Francisco Bay Area, where it has already mapped the roads. It’s like cruise control on steroids — with the ability to follow the road as it curves, stay in a lane, accelerate and brake according to traffic. The Cruise kit costs $10,000, and 50 will be sold in the first run.

So basically, Cruise is a fancy way for venture capitalists to commute.

But more than that, it’s an incremental step toward a self-driving car, rather than the more audacious fully baked version that Google is working on — where it is actually building cars from scratch. Google has said it expects mass-market versions of its cars to arrive between 2017 and 2020.

The Cruise team openly credits the progress Google has already driven in the three big areas for autonomous vehicles: Popular awareness, regulatory approval and technology. “We have to get right behind them and draft off their progress,” said Cruise operations head Daniel Kan.

Re/code has been tracking Vogt’s progress at Cruise over the past few months, taking multiple rides in the company’s prototype vehicles as it ironed out the kinks (hopping in self-driving cars has kind of become my thing). Last week, we attended a test-drive event in Alameda, Calif., where a Cruise-modified car danced between cones, braked around a tight turn, and then accelerated on a straightaway. Check out a video of that here:

If they want to abide by the law, people in the drivers’ seats of Cruise vehicles will still have to keep their eyes on the road. They’ll just be offloading some of their mental focus to the car. They don’t need to keep their hands on the steering wheel (unless they are in a state that requires one hand on the wheel).

“Most people, when they’re driving or commuting in stop-and-go traffic, they’re sitting behind the wheel fuming at the car in front of them, wishing it would go faster,” said Vogt. “But when you have this system, you’re somehow a little more disconnected, and you don’t quite feel that pain and stress, so hopefully that won’t come home with you and overflow into the rest of your life.”

Cruise doesn’t just want to keep cars in their lanes on highways in the Bay Area, nor just to work with Audis. It does have big ambitions, eventually. What’s notable is that it’s taking an incremental approach. Vogt founded the company just seven months ago and put together a working demo in March as part of one of the more aggressive entrants in the Y Combinator startup education program.

“We plan to do much more,” Vogt said. “The world is better off with this thing than without it.”

If Cruise has a time advantage, it won’t be for long. The fantasy world of self-driving cars is quickly turning into reality, and this is not just a Silicon Valley story. Based on reduced accidents, relieved congestion, increased productivity and other factors, Morgan Stanley estimated late last year that self-driving cars would save the world $5.6 trillion per year, and that fully autonomous cars will be available before the end of the decade.

Veteran car-vision company Mobileye, based in the Netherlands, says it plans to launch hands-free driving at highway speeds with multiple car partners “in the 2016 time frame.” A company called Induct this year started selling a little bus called the Navia that can go 12 miles per hour. Many car makers around the world already offer advanced adaptive cruise-control systems using radar, as well as nifty parallel-parking helpers. For instance, an existing Acura upgrade helps keep cars centered in their lanes, though it only works in short spurts, due to safety concerns. Cadillac showed off something very similar to Cruise in 2013 — it even calls the feature “Super Cruise” — and said it would be ready later this decade.

Vogt claims Cruise will beat smaller and larger rivals to the market at a relatively affordable price. “To my knowledge, there is no driving system on the market that can drive your car hands-free from 0 to 80 miles per hour on highways,” he said. “Plenty of demos have been shown at CES, but if you go to a car dealership today looking for this technology, you’ll be disappointed.” Cruise will also send software updates over the air, so additional roads and features will be beamed down automatically to customers.

Like Google’s self-driving cars, Cruise depends on building its own more detailed maps of roads. Its components include a mounted roof pod with cameras and radar — not the $75,000 spinning lasers found in all of Google’s cars — a computer in the trunk, and hidden actuators that control steering, braking and acceleration. Cruise does not have a formal relationship with Audi, and Vogt believes that the “minimally invasive” system would not void an Audi owner’s warranty.

As Vogt puts it, the reason a startup can compete in this space is because of the intersecting trends of cheap and good sensors and cheap and good computers. Because many of the big-picture pieces are already laid out, Vogt and his team can spend their days tweaking their algorithms to do things like making steering smoother.

“We are a new company, and I understand some people are not going to trust us in the same way they would trust a car company that’s been around for 100 years,” said Vogt.

He added, “This is not the holy grail. But it’s a first step, and I think some people will really enjoy it.”

Self-Driving Cars Coverage




3 comments
krkeegan
krkeegan

Interesting, there appears to be a lot of doubt about the ability of technology in the comments of a tech blog.


I am wondering how these vehicles handle merging on a congested highway?  I have no problem believing that they can do a perfect job on a freeway at normal speeds.  But how "aggressive" can the vehicles become?

hockey queen
hockey queen

I saw a video of the car swerving into the left lane to avoid a person in the lane the car was in, so, how will it handle the scenario of person in car's lane and oncoming car coming in another direction? Will it swerve to avoid person and cause head on in the other lane?

In just the last month I have had to avoid a barrel on the highway in a lane, avoid a spare tire that was rolling down the highway going in same direction of traffic along to the right of the cars and the tire ended up veering in front of our lanes and ended up crossing the median and crashing into the windshield of highway traffic going in the opposite direction. If I swerved I would have side-swiped other cars. Sometimes swerving just will cause more trouble, sometimes you have to slow down or completely stop... sometimes the object is not where it will be picked up by cameras and the computer won't be able to know how to react.... too many decisions to process to decide what the best decision is.

How about two lane highway, you are in the left or right lane and the car to your right or left decides to try to change lanes and doesn't see your car, effectively almost sideswiping you? Will the computer car realize and blow on the horn? or will it slam on breaks or accelerate to avoid? and if it slams or accelerates will it cause another accident with car ahead or behind? or will it swerve to avoid? and maybe run off the road...

I just think there are too many decisions to make and how do we know that the computer designers programming will be able to program the car to make the best/optimal decision if we aren't sure that all options are even perceived/picked up by the computer program?

jsmitty
jsmitty

Why does anyone think driverless cars are a good thing?  Car companies are struggling to make cars that are safe and without glitches already.  Do you really think anyone will be able to make a driverless car without glitches?  I foresee senate hearings and banning driverless cars after the first death.  Here are some questions I have:


1. What happens if you're in a construction zone where roads are often changed?

2. What happens if there are small objects driver sees in the road, such as nails, screws, etc. that you want to avoid?  Will driverless car avoid them?

3. What happens when you decide to change your destination at the last minute?  How will the driverless car know that?

4. What happens when someone uses hand motions to wave you ahead?  How will driverless car know?


These are just issues off the top of my head.  I'm sure there are thousands of other issues.  Everyone talks as if driverless cars will result in safer roads.  I predict they will result in chaos on the road.

Follow

Get every new post delivered to your Inbox.

Join 279,222 other followers