Cruise, a unit of General Motors with investment from Honda, revealed a new “not a concept” self-driving vehicle last week. The vehicle is custom designed for self driving, with large sliding doors, electric drive and face-to-face seating. Cruise CEO Dan Ammann outlined their vision for the vehicle and it has much in common with the robotaxi vision of many companies, including the one that Larry Burns and I promoted at Google/Waymo. But the most interesting similarity is to Zoox, the billion-dollar funded Silicon Valley robocar startup.
The Cruise Origin, which is “not a concept” features face to face seating, no controls for riders, … [+]
Cruise
Zoox has kept a low profile for much of its existence. Because I explored Zoox’s vision with its founder, Tim Kentley-Klay, extensively before its founding, I have a few special insights into their vision, which involves creating a vehicle custom designed for self-driving taxi service from the ground up. It is worth contrasting the core visions, prior to this, of these players:
- Waymo: Build a superb self-driving system, then operate a robotaxi fleet by ordering moderately modified versions of existing car designs (such as the Chrysler Pacifica minivan.) Waymo’s view is that building a self-driving system is more than hard enough, so it’s better to exploit the virtues of mature, existing vehicles and modify them for self-driving.
- Zoox: Having a custom vehicle design offers several advantages (described below) and they are large enough to make it worthwhile to try to both develop the self-driving system and a whole new vehicle design, effectively being a self-driving company and a radical new car company at once. At the same time, R&D on the self-drive system is done with modified standard cars.
- Cruise: They’re a self-driving car system company, developing on existing car platforms. But now, because they’re owned by a car company like GM, they can also develop a custom car for self-driving. While any radical new car is hard, they don’t have to create a new car company.
- Other OEMs: They make cars, and are building self-driving systems to add to their cars with minor modifications.
- Argo (Ford) released their own thoughts about customizing vehicles some time ago, but kept to more mild modifications.
- Tesla: A radical new car company which has designed its production car to include a sensor suite that they hope is enough for self-driving. They feel that their fleet gives them an advantage in trying to develop a self-driving system, and that as such they can do it better, with far less in the way of sensors.
- Aurora and many other startups: They just focus on the software or sensor hardware, and hope to partner with existing OEMs when the time comes.
- Lyft/Uber/Didi/etc.: They are taxi network companies who already have the taxi business. They want to also have a self-driving system to deploy on cars or license out to partners who enter their taxi network.
There are more approaches than this, but it is interesting to see Cruise go head to head with Zoox. Cruise’s Origin car is still in early design, there are no hints of when it might enter production, but they insist it is not just a concept. Much earlier, Cruise showed off a design for a modified Chevy Bolt with no pedals or wheel, and said they were going into production on that, but little more has been revealed. Zoox’s custom vehicle is still in development, and has yet to drive on public roads. Zoox has been working on it much longer, but GM has much more experience designing cars and putting them into production.
This teaser of Zoox’s prototype design shows 4 wheel steering, allowing the car to be vastly more … [+]
Zoox
Cruise and Zoox both feature designs with face to face seating and forward and back symmetry. It is likely the Cruise Origin, like the Zoox, can go forward or backward equally. As custom designs, both will be built to make sensor placement easy and effective. Both are designed for passenger comfort, ease of entry and exit, and long vehicle life. Long vehicle life an interesting factor in the taxi business. With taxis serving 60,000 or so miles per year, current cars only survive about 5 years of taxi service. With a large portion of the cost of a robotaxi service being depreciation on the vehicle, long life can offer economic advantages. On the other hand, if vehicles go obsolete quickly, you may not want 10-15 years of life. Cruise says they account for that by using modular designs that let them replace sensors or other components – this is a good design approach. (The interiors will also need upgrading, probably every 50,000 miles or so.)
The Cruise Origin Is meant for sharing, by which I mean ridepooling, not simply taxi service. That can be efficient, but in many cases having 1-2 person vehicles can be better than trying to group unrelated riders in a larger car, except at rush hour when road space is at a premium. For privacy and safety, people may also wish that ridepooling vehicles have isolated compartments which demands a different door configuration from the Cruise Origin.
There has been much debate over whether Zoox’s bet is smart or crazy. It’s definitely harder to do it all, though there are some synergies that make some parts of it easier. Given a choice, riders will probably prefer a custom designed taxi vehicle to a repurposed passenger car, as the former will offer more comfort, room and convenience. Unfortunately, head to head competition in taxi service is unlikely in the early years of the robotaxi business as each vendor will be likely to prefer opening up virgin territory in a new city than deliberately going into head to head competition.
Very early designs from Zoox show concepts of face to face seating, symmetrical design, no … [+]
Zoox
For Zoox, there is a definite challenge in completing with GM on this. GM is much more experienced and capable at making cars. Zoox’s core advantage there is its radical, “non car company” thinking. This is not to be underestimated – that same advantage at Tesla has helped it dominate its price class against the big boys.
Why a custom car?
What are the key advantages of building your own custom car, as Zoox, Cruise and most shuttle developers (like Local Motors and others) are doing? Waymo even created a custom car, known as the Firefly, for their 3rd prototype, before switching back to designs based on existing cars.
1. Today’s cars are designed around the driver, and almost everything in the car serves the comfort and convenience of the driver. Many of those resulting decisions use a lot of space and constrain the design. Freed of this, you can do a lot more for the passengers.
2. Taxis, in particular, can be custom designed to the areas they will service. In particular, taxis that won’t go on highways or go fast do not need to be as aerodynamic, allowing a lot more choice in design, and a lot more room. This can include vehicles where you can stand. You can also have narrow vehicles for 1-2 people which can take far less room on the road, be easier to park/store, and which can even go between lanes or down special right of way only for narrow vehicles.
3. You can design the vehicle to make best and most economical use of sensors, making sure they get the necessary view of surroundings.
4. For people who want to socialize, seating can be face to face (though this takes more space) or it can also be easier to design private compartments for strangers who will ride together.
5. Vehicles can be bidirectional, so they never have to “back up” and can have a simpler time picking up and dropping off passengers.
6. While some passengers will want a view, for many, you are freed from having a big, single pane windshield for those who don’t care – or you can have a better one.
7. While crumple zones and other passive safety remain necessary (though less so on vehicles that only go at slower speeds) you get more options on how to install this. Indeed, rear facing passengers are generally safer in any forward collision, which are generally the highest energy crashes.
8. Door designs can make it easier to get in and out.
9. For highway cars, designs can be more aerodynamic and efficient, particularly for passengers who will ride backwards. (The most aerodynamic shape looks like a teardrop – though the half teardrop is a good compromise for a ground vehicle.) This will be tallest near the middle front and tapered down at the back.
10. New designs permit all sorts of new ideas that, while as yet untested, never were even considered for classic 4-5 passenger forward facing, human driven cars. We may not yet know what the best advantage of having a clean design sheet will be. Many prototypes have been built – some clever, some silly, with features like coupling, tilting, merging and more. This should continue.
These are great advantages, but they come with a huge cost compared to the sure bet of relying on existing car design knowledge. Others are betting that you can get a good car going with existing designs, and then go wild later. Only the future will reveal the right bet.