Driverless Conference sparks autonomous car development analysis
GPS and GNSS have changed the world. Of that there can be no doubt. But in terms of sheer change, both qualitative and quantitative — we ain’t seen nothing yet.
We now witness the creation of an industry. This will be very disruptive. We’ve had change instituted by GNSS; we know what that looks like. We haven’t yet seen a true revolution.This is something entirely new, and there are many things about which we don’t yet have a clue .
What happens to that great American institution, the private car? The relationship between the individual and its four-wheeled extension?
And on the industrial side, do automakers disappear as OEMs — do they become Tier 1 suppliers to Google, Uber and Lyft?
Because of the massive impact of this particular rollout of GNSS-enabled capabilities, I am devoting this issue of the GNSS Design & Test e-newsletter to it, even though it is not in itself a system in space. It is the most radical transformation of life on Earth we have seen, driven by our systems in space.
The following are notes jotted during the Driverless Conference, March 23 in San Francisco.
“In the early 90s, when I was part of a government ride-sharing initiative, we used to talk about these new portable devices bringing data communication into … wherever we go. Now they’re here, and they’re well established. Very soon, this is going to change things, and enable many of the things we’ve only talked and dreamed about so far.” Thus spoke Steve Wollenberg of Automatiks, opening the conference.
“We’re at the confluence of great technological developments. We’re seeing great acceleration of all of them.”
Virtually all the speakers, all these driverless enthusiasts, really love cars. Some own up to collecting them, having multiples in their home garage(s). A bit wistfully, Wollenberg foresaw the new control technology taking over public roadways. “In ten years, racetracks may be the only place where you’re allowed to drive your own vehicle.”
Ride Share. “Four years is the entire lifetime of the ridesharing industry,” said Emily Castor of Lyft, who by virtue of her tenure there for that period, can be termed an industry veteran.
“We’ve seen a full-about turn in the regulatory environment. We see ride-sharing as the stepping stone to a world in which people no longer drive vehicles. Getting an autonomous vehicle on demand through a shared network will be much easier and cheaper than owning a private vehicle.”
Lyft talked with General Motors last year, and found a shared vision of shared use.
Amitai Bin-Nun from Securing America’s Future Energy (SAFE), a non-partisan advocacy organization with business leadership, introduced his organization’s broad mission: reducing U.S. petroleum dependence. Instability in parts of the world is fueled by petroleum dependence.
“This is a hard process. It takes a long time to overturn an established system.” A key obstacle is the lack of compelling new consumer experience, currently. Using connected and autonomous vehicles in a ride-sharing network is an opportunity to get this new experience, and drive the transformative process. Re-order the transportation system.
Mariel Devisa of Travelers Insurance announced that Travelers has launched a ride-share insurance product, live now in 16 states.
In two fairly conservative industries — automotive and insurance — with long-established partners and practices, the efforts to move and change are, frankly, surprising and faster than anticipated, according to moderator Wollenberg. “It’s a fun time.”
Freight and Fleets. Steve Boyd of Peloton made the case that trucking fleets can serve a critical role in pushing the technology forward, because some segments of the transportation industry move faster than others. Getting state approvals without having to go federal is the route pursued now, in terms of full-scale roadtesting of autonomous driving. That will enable early adoption heading into commercial pathways: freight-truck platooning and drafting. Volvo, Intel, Nokia, Denso, UPS and a number of other companies are closely involved.
Boyd announced a set of fleet trials this year, starting in Texas, “a very truck-friendly state.” Legislative approval for trials has passed or is pending in several other states, as many as a dozen. Prospective customers are already lined up in the freight space.
In Europe, an April 6 EU Platooning Challenge will take place in Rotterdam. The Netherlands is leading the EU in the current cycle to approve truck platooning by early 2018.
There’s “a platooning gap” developing between the U.S. and Europe, according to Boyd. Silicon Valley may lead on the technology, but if this is not matched by activity on the regulatory side, it will lose out to other areas that aggressively pursue approvals as well as technology.
Traditionally, the automotive and trucking OEM industries have been very competitive, but now they are seeing the necessity to collaborate to push the policy side forward. This is happening in the insurance industry, too. Competition will certainly still be there, but to enable vehicle-to-vehicle communication a broad measure of collaboration will be necessary.
The road environment today is very imperfect, with many thousands of fatalities and countless more serious injuries. Trucks drive too close together. Highway safety needs innovation and regulatory change in order to improve.
The Long Vision. An autonomous car can’t count on the ability of the driver to retake control of the vehicle in 5 or 10 seconds. So the vehicle needs to be able to take care of itself — fully. Therefore, an evolutionary approach to installing autonomous capabilities may not work.
Some initiatives, however, continue to bring services into the vehicle one by one, gradually. How engaged will the driver be, and in what timeframe? There’s debate, and a shift in thinking may currently be underway.
Traditionally, a 5- to 7-year product cycle in automotive starts when new features are introduced in upmarket vehicles. Examples: adaptive cruise control (to follow the car in front of you at a set distance), lane-keeping assistance. Gradually, these new features are installed in lower price-point models until they become standard throughout the line. With multiple products and product cycles, it will thus take multiple decades. 220 million vehicles are owned by households. An integrative approach to autonomy will take a long, long time.
There is a rising tide for autonomy may take a different approach: autonomy first, that is, full autonomy will take over the vehicle — and as many vehicles as possible.
(Something that no one has mentioned but I can’t help thinking: Given the longstanding and extremely virulent controversy in this country over private gun ownership… What does this bode for something shaping up as a massive social, structural change, not just a new technological wrinkle? What is more American than a gun? A car.
If you thought the Internet, or smartphones, or for heavensakes even GPS/GNSS have radically altered the world — again, we ain’t seen nothin’ yet.)
the problem is that the self-driving cars depend on a perfect world,the ability to communicate with both infrastructure and other vehicles, and on everything being predictable. But the world is not perfect, ice patches are an obvious example of that. The cost and effort to provide the smart infrastructure that these vehicles need is both unreasonable and quite prohibitive, and not likely to be more than 80% active at any time. Communication with all other vehicles in the area is a big challenge, and verifying that they are providing correct information is probably not possible. And we know that many aspects of this world are far from predictable. So from a technical point of view the success of autonomous vehicles is not predicted.
But the very worst aspect is the end of personal freedom that is provided by personal vehicles under a driver’s control. The very next step with the autonomous vehicles, already in constant communication with a master control system, will be the ability of those in charge to say “NO”, and prevent the vehicle from moving. The concepts predicted in Orwell’s book “1984” were way off in their timing but not in their content. Disallowing driving during periods when those in charge don’t want driving will certainly be within the basic control scheme of all of these vehicles. Are we willing to accept that?
Sounds like an upbeat conference. Wish I could have been there.
Thanks for the synopsis.
For us GNSS folks, we know that GNSS technology is a key element, but not the main component of self driving. Its a combination of sensors, vision, communication, and a huge amount of software/AI/machine learning. Thanks to Moore’s law, we can get this serious amount of computing at a good price. We see the Qualcomms/Broadcoms making a play at the automotive compute side with multi CPU core offering. We might also see low enough cost RTK as uBlox enters this market. But in the end, the always-required safety margin will probably not be dependent on the availability of cm level GNSS positioning: rather, the accuracy will be derived from infrequent high accuracy updates coupled with an awesome dead reckoning technology. Otherwise, you couldn’t have deliveries in the urban canyon without slowing down the old cars with humans inside.
Here is another question about that utopian ride sharing vision: How can a shared vehicle be an individuals personal space? My vehicle contains a lot of things that stay in it all the time, starting with driving glasses, gloves, and other such accessories. A ride shared vehicle would be far closer to riding a city bus, in that everything wanted on the ride must be provided by the rider, and then removed by the rider at the rides destination.
My guess is that the proponents of this concept are some of those who mandate that there should never be any item in a car except for the OEM provided things and items to be carried on that particular trip.