In the Cadillac video below, they explain the system as a combination of ACC, lane-keeping and GPS. This is similar to the other announced plans from many other car companies, including Mercedes, BMW, VW/Audi and others. The use of GPS suggests the car may also use map information, which is not known to be used by the other announced products, but is heavily used by Google and the various eyes-free projects.
It is pure speculation, but perhaps they are building maps of where the lane markers are reliable enough and where they have faded out so that they can refuse to super-cruise when approaching those zones. They might also use the GPS to assure you super-cruise only on the highway or other limited areas.
In the video, which shows a demo at about the 1:10 mark, they are driving on a test track, and always next to a blue line along the lane markers. Obviously a real product could not depend on special lane striping if it wants to be broadly usable, but this may assist them in testing their system with confidence. (ie. compare what their lane-finder detects to what an independent system that tracks the blue line detects.)
GM has had various self-driving projects, including the futuristic EN-V and the sponsorship of BOSS in the Darpa Urban Challenge. The Cadillac brand is well positioned. Self-driving is initially going to be a luxury feature, but companies that sell sporty performance cars don’t want to detract from their image as selling a fun driving experience. A pure luxury brand like Cadillac does not have as much of that problem as BMW and Mercedes have. At the same time, the video insists that they don’t want to take away from driving.
Today Google released a new 3 minute video highlighting advanced self-driving car use. Here I embed the video, discussion below includes some minor spoilers on surprises in the video. I’m pleased to see this released as I had a minor & peripheral role in the planning of it, but the team has done a great job on this project.
This video includes active operation of the vehicle on not just ordinary streets, by private parking lots for door to door transportation. You can click on it to see it in HD directly on Youtube. read more »
You may not know the name of Continental, but they are a major supplier of components to the big automakers. A story in the Detroit Free Press details their latest project in autonomous driving. This is a VW Passat using radar Automatic-Cruise-Control combined with lane-keeping, similar to projects announced by Mercedes and VW/Audi itself. The story has a video showing the screen of the car displaying its lane-keeping. The car also has side radars to track vehicles or barriers to the left and right, according to the story. It’s aimed at stop-and-go traffic and empty highway. If it’s like the other products it requires constant human supervision, as it is not safe to not look at the road in case the lane markers vanish or other unexpected problems occur.
They claim they have done 6500 miles in Michigan, and that soon they will have the 10,000 needed for a testing licence in Nevada. The new Nevada law allows developers of robocars to test on Nevada highways once they have shown 10,000 miles on a test track or in another state, and under special testing rules. (The Google cars have over 200,000 miles in California and Nevada.)
The Nevada regulations specifically exempt vehicles which require full time human supervision, so in theory they don’t need a Nevada testing licence if this is such a vehicle. If it is planned to operate without such supervision it needs the licence and is more advanced that the other systems of this type.
An interesting note about the photo, credited to Conti — if this car actually does qualify as an autonomous car in Nevada, then that picture of the car robo-driving in Las Vegas presumably was taken before the regulations came into effect.
A recent article on bicycles and pedestrians in the robocar world appears at the Greater Washington web site, which has taken an interest in robocar topics. In particular they are concerned about the vision of a reservation-based intersection, which does not use traffic signals. These designs from U of Texas got a lot of press in the last few weeks after a presentation at AAAS, but they’ve been around for years and I have a number of links to them. What’s new is that the coming of robocars makes them seem more practical.
In a reservation based intersection, the computer handling the intersection hands out slots to cross the intersection. The slots are moving boxes that you have reserved, and you cross in them. The computer hands out the boxes so they never hit one another. The simulated result at first would scare people to death but over time they might trust it. However, it requires that every car on the road have automatic operation, since deviation from your reserved box does indeed mean serious risk. Human judgement just would not cut it here. As such, intersections like this are a long, long way away.
Closer, I think, is the concept of reservation based roads. These are road segments which hand out long term slots, such as “You can drive this block between 8:30 and 9am.” The road only hands out as many slots as it can handle, but does not try to schedule the cars down to the square foot-second. In such a system, as you approach that block on your trip, you would refine and correct the initial reservation, so that by the time you are a minute away, your window is just a few minutes. If roads can do this they can assure, well in advance, that they never get more cars on them than they can handle, and this reduces the odds that traffic will collapse due to congestion. The biggest cause of congestion is basic excess of demand over supply — accidents are the #2 cause.
Such a system can also handle human driven cars. Those cars are a bit less predictable and need wider reservation windows. They also will eventually need more space on the road, since robocars will eventually start packing themselves closer together once they are common enough to do that. Half-width robocars will commonly pair up in a lane with other half-width vehicles.
So what about the bicycles? It will be daunting for them. If there is a bike lane, that’s great of course. And at “bike rush hour” we can even make sure “parked” robocars get out of the way to make a bike lane if that’s what we want. (We may want another car lane even more.) Otherwise a virtual bike lane can be made if the bikes have to ride with the traffic.
Bikes do present a safety issue to be sure. In the worst case situation, a cyclist can fall off their bike and stop immediately, lying in the road. A vehicle following a bike has to leave enough space to assure they can stop before that, including reaction time. Reaction time should be better for robocars than for humans. Humans don’t leave enough space right now. We leave even less space behind cars because cars actually can’t stop super fast, and you brake with them. and if you hit them at slow speeds it’s “tolerable” — nobody will be seriously hurt. Hitting a cyclist or pedestrian at slow speeds can mean death.
(Head-on collisions are a different matter and they can cause great mayhem. I believe that moving mostly to one-way streets is the best solution to the problem of head-ons, and with robocars, the inconvenience of one-way streets can be greatly reduced.)
Robocars should end up much better at spotting cyclists than humans are, because robocar vision is 360 degrees and in 3-D. There are no blind spots in a robocar system and it’s always paying attention in all directions. The only negative in spotting them is their small size. A bike that appears out of nowhere from behind an obstruction is always at risk to both robocars and human drivers. Robocars will work very hard to not hit cyclists, and in fact in the future street that’s 100% robocar, a cyclists should feel pretty safe, and could even abuse the system, weaving back and forth and causing jolts for the passengers around the bike.
On the plus side, robocars might enable two things. The first would be the creation of dedicated lanes, paths and even elevated guideways for use by both bicycles and narrow lightweight robocar trikes. I anticipate these lightweight vehicles will become very common, as they are the most efficient vehicle for short urban trips. Because they are light and small, it’s vastly cheaper to build dedicated pathways and elevated guideways for them. These guideways could be made open to bikes if there are passing zones, since the robocars would sustain higher speeds. (We have not yet convinced many US cities to dedicate a lot of space and money to bike-only paths, otherwise that would be obviously better for bikes.) Robocar only lanes offer a cheap way to increase road capacity and offer ultralight robocar users a faster, zero-congestion trip in the busiest areas, and thus make a lot of sense for cities. The bang/buck is as high as it can get in transportation development, and it encourages green transportation, as these trikes use less energy/person than transit systems do.
Another interesting development might be the bike-bot. As I envision it, this is a very small robot that’s able to clamp onto a bicycle and move the bike from place to place, using the bicycle’s wheels as well as its own. This could offer a world of “bikes on demand.” No matter where you are, you could summon up a bicycle in a short time, and drop it anywhere. (At your destination, you would insert the bike into a bike-bot that sent itself there ahead of your arrival, and the bike-bot would take the bike to its next rider.) This could make bicycle use very convenient, and would be good, efficient exercise for all who need it.
I also suspect that we’ll see ultralight robocars that feature pedals. With the pedals, the rider would have the option of exercising and their energy would also go into powering the vehicle. The commute is a good time to exercise and watch videos or read. Not as much fun as recreational cycling, but more pleasant in other ways that cycle-commuting.
In the more distant future, when all cars are robocars, we will begin to see the conflict between the cars and the bikes and pedestrians described in the article cited above. The author is right that putting pedestrians on elevated bridges is not a good answer, and forcing bikes off valuable road is not good either. In an idealized robocar road, which has no parked cars on the side, and just many lanes of one-way traffic, the presence of the cyclist does use up a lot more road capacity per person than the cars do. We’re a long way from that idealized capacity, but should we come to depend on it, we might see pressure to push the bikes away, or charge them or the amount of square-foot-seconds of road they use. That will be a political decision, where we may decide many decades from now that to encourage cycling, it’s worth subsidizing it a bit.
The state of Nevada today approved regulations for self-driving cars in the state. Last year, Nevada passed a law outlining the path to these regulations, and their DMV has been working in consultation with Google, car makers and other parties to write them down. Today they were approved, allowing testing, certification and — someday — operation of vehicles in the state. Other laws are in consideration in other states inspired by the Nevada move. This is, frankly, much sooner than I anticipated.
In other news, a junker car race known as “24 hours of LeMons” (completely unrelated to Le Mans) has announced that self-driving cars may enter and are exempt from the normal requirement that cars cost no more than $500. The “X cedingly bad idea prize” of a million nickels (not quite as good as X prize purses of $10 million) probably won’t get too many takers at first. This race has a sense of humour but I’m not sure too many folks would risk their expensive autonomous car on that track or feel it safe enough to drive with crazy amateur racing drivers. I suspect they don’t really mean it and just wanted to issue a press release, but it will be fun when robocar technology is common enough that garage tinkerers on low budgets can enter races like this.
A new group has sprung up in the valley around the concept of the open source automobile. I will be speaking at their meetup, which was going to be at Hacker Dojo and has moved to Intel’s auditorium. Looks like a good crowd is signed up. Sorry, no deep Google secrets but there will be a video featuring the car and many visions of the future.
Update: When I first wrote this, I was under the mistaken belief that Better Place only swapped one type of battery module. At present they only support one, but their swap stations are designed to support up to six kinds, as long as they can be loaded and unloaded from below.
Recently, electric car battery-swap company Better Place announced delivery of their first cars in Israel. Israel is a country of small size where it makes sense to deploy a technology like this with a chicken and egg problem. They hope to have enough battery swap stations that people feel they can drive an electric car and refuel it as quickly and conveniently as a gasoline buggy.
I remain skeptical about battery swap for electric cars, but think robocars solve many of those problems. Here’s why:
To have a workable battery swap system, you need to standardize the battery module, ideally having just one or two form factors and electrical characteristics. Just one to start, in fact. This has many downsides:
A large part of the innovation in electric cars today is in the batteries. A big part of what Tesla did was their new cooling system. Designers all want to be able to play with chemistries, voltage, controllers and more. They might give up playing with size and placement but not those things.
It’s still an issue to not be able to vary size and shape of the battery, at least for people wanting to build cars of unusual shape.
A large part of the cost of an electric car today is the battery. With the swap system, you can’t buy the battery with the car. You get whatever battery you are swapped. That’s good in some ways, but eliminates open market competition on these systems because there is only one buyer — the swap company.
Swap machines are expensive, take land and still take five minutes, arrival to departure. That’s almost as quick as filling up with gas, but a typical gas station has 8 pumps, and some have many more. If there are several cars in line at a single swap station, you’re in for a serious wait.
On the plus side, people actually need swaps more rarely than they imagine. A 70-100 mile range car will hardly ever feel the need for a swap — in many ways the availability of the swap makes you feel more comfortable about the car, even if you rarely use it. It’s better than the level 3 charge which can damage the battery and still takes 15-20 minutes.
I think robocars (cars able to move while empty to the swap station) solve many of these problems. They solve them because while robocars (particularly those operating as taxis) need to run all day and thus want to swap batteries, the cars can move to the swap station on their own, when they are not serving somebody.
There is much less need to standardize, though it does help. Your car simply goes to a swap station that has its type of battery available.
While it wastes energy and a little time, it doesn’t bother the robot to have to go a few miles to find such a station. You don’t need one on every popular route.
The robot can schedule an appointment for a swap if need be. Not that it really minds waiting a lot, unless it has a job to do. But with a scheduled swap it might even do one while carrying a passenger, if it happens to be passing a swap station and can book a no-wait appointment for the time it will be passing, and the passenger doesn’t mind the 3 minute stop.
Most typically, this will be used by taxi fleets. Each taxi fleet can have their own swap station, for the type of battery cases they like. They can program their taxis to take jobs that bring them closer to the swap station when they will be running low. You can get buy with just one swap station for the whole fleet, or perhaps just a few. The taxi fleet can have a mixture of cars of different swap types and cars without swap ability. The latter can’t run all day and must spend time in charging stations as planned.
With robocars, you can solve range problems not by swaping the battery, but by swapping the car. If you have a car that is running low, it can stop in a convenient lot to have you switch quickly to a taxi with lots of charge. Then the car you were in can head off for charge or swap in no paticular rush.
By allowing lots of types of battery form factors and swap stations, you allow innovation and competition in these areas, which in the long run is a win for the customer. Anything that blocks competition may sound good at first but quickly bogs things down.
Now I still want to credit Better Place for working to solve the range and range anxiety problems of electric cars. There will still be competition because I don’t expect all electric car vendors to want to be compatible with their system. But I think their technology comes into its own best when the cars can worry about the swap rather than the drivers.
BMW has spoken in the past of their Highly Automated Driving project. A new video about ConnectedDrive Connect goes into more details. While BMW insists this is just a research project and won’t be in their cars for a long time, they also are expected to soon produce a traffic jam autopilot to compete with the offerings from Mercedes and Audi.
The approach in this car is a bit different. They have a small LIDAR on the side of the vehicle, which presumably is there to find the lane markers, and a forward facing camera behind the rearview mirror. The other systems are mostly using cameras for their lane finding. They also have ultrasonics, which are typically used for automated parking and some blind-spot detection. LIDARs are currently much more expensive than cameras, but this is a concept car for now. (More complete robocars tend to use fairly advanced LIDARs.)
The 5,000km was reported back in September This comes after reports that their track trainer had done about 12,000 miles in a 3-series on the track. There are more details on this system at MIT Tech Review.
I’ve talked before about Traffic by Tom Vanderbilt. It’s an interesting journey into the world of traffic psychology and engineering, and has one chapter on robocars. That chapter was based on seeing Junior at Stanford, and things have come a long way since then.
Friday, Wired released Let the Robot Drive, a decent length article on the Google car (he came for a test ride) and various other projects, including more about the Mercedes 6-D vision system. For those hungry for Google Car news it’s a good place to go.
There’s also an article by John Markoff about the SCU legal conference I wrote about on Saturday, which finishes off quoting a joke I said there.
At the same time a different long article with a more futurist bent appears at the World Future Society web site by Thomas Frey.
Yesterday I attended the conference on the legal implications of robocars put on by Santa Clara University Law Review. It was a well done conference, with some real content from a varied group of regulators and legal scholars, a sign of how real the robocar world has become.
After a technology introduction where Sven Bieker of Stanford outlined the challenges he saw which put fully autonomous robocars 2 decades away, the first session was on civil liability. The short message was that based on a number of related cases from the past, it will be hard for manufacturers to avoid liability for any safety problems with their robocars, even when the systems were built to provide the highest statistical safety result if it traded off one type of safety for another.
In general when robocars come up as a subject of discussion in web threads, I frequently see “Who will be liable in a crash” as the first question. I think it’s a largely unimportant question for two reasons. First of all, when the technology is new, there is no question that any lawsuit over any incident involving the cars will include the vendor as the defendant, in many cases with justifiable reasons, but even if there is no easily seen reason why. So potential vendors can’t expect to not plan for liablity.
But most of all, the reality is that in the end, the cost of accidents is borne by car buyers. Normally, they do it by buying insurance. But if the accidents are deemed the fault of the vehicle maker, this cost goes into the price of the car, and is paid for by the vehicle maker’s insurance or self-insurance. It’s just a question of figuring out how the vehicle buyer will pay, and the market should be capable of that (though see below.)
No, the big question in my mind is whether the liabilty assigned in any lawsuit will be significantly greater than it is in ordinary collisions where human error is at fault, because of punitive damages. The cost of collisions is well established and understood, and there is a large industry to manage it, and if robocars are, as planned, safer, then that cost goes down, and that means savings for the car buyer and for society — a big win for all. However, if the cost per collision is much higher even though the number of collisions drops, this can impede the ability of a robocar industry to innovate or even to exist and save those injuries.
Unfortunately, some liability history points to the latter scenario, though it is possible for statutes to modify this.
All cars must have insurance today, and it is normally bought by the car owner/driver, and covers almost all accidents. If a safe robocar is delivered, the lower rate of accidents should mean cheaper insurance. Alas, in California, one of the world’s largest insurance markets and home to robocar labs at Google, Stanford/VW and others, it’s not so simple. California’s Proposition 103 demanded that any insurance policy’s price must be based on weighted factors, and the top 3 weighted factors must be, in order, driving record, number of miles driven and number of years of experience. Other factors like the type of car you have — ie. if you have a robocar — must be weighted lower. So this law makes it very hard to give very cheap insurance for a robocar, and makes it close to impossible to do it for somebody who is getting the robocar precisely because they are a bad driver and have found themselves facing expensive insurance if they keep driving.
Because Prop 103 is a ballot proposition, it can’t easily be superseded by the legislature. It takes a 2/3 vote and a court agreeing the change matches the intent of the original ballot proposition. One would hope the courts would agree that cheaper insurance to encourage safer cars would match the voter intent, but this is a challenge.
Kevin Vincent and Steve Wood from NHTSA, which writes the federal motor vehicle safety standards and does the crash tests, among other things, spoke and gave a fairly robocar-positive message. They are working to assure the vehicles operate safely on the roads, but do appear to be fully aware of how they are a new technology under new rules, and they don’t want to impede adoption by overregulating. They are working out plans to test vehicles in simulation, on test tracks and on the road to rate for safety and seem to have a generally forward-thinking plan.
Local and criminal laws
The session on criminal laws centered more on the traffic code (which isn’t really criminal law) and the fact it varies a lot from state to state. Indeed, any robocar that wants to operate in multiple states will have to deal with this, though fortunately there is a federal standard on traffic controls (signs and lights) to rely on. Some global standards are a concern — the Geneva convention on traffic laws requires every car has a driver who is in control of the vehicle. However, I think that governments will be able to quickly see — if they want to — that these are laws in need of updating. Some precedent in drunk driving can create problems — people have been convicted of DUI for being in their car, drunk, with the keys in their pocket, because they had clear intent to drive drunk. However, one would hope the posession of a robocar (of the sort that does not need human manual driving) would express an entirely different intent to the law. While the car might have an emergency manual override, as is likely for other reasons, one would hope it is the drunk’s responsiblity if they use it, not the car’s responsiblity for having it there.
The session on privacy left me a little wanting. There was no addressing the issue that cars will have cameras on them, recording the whole world once they get frequent enough. Most of the concern was on the logs of all your movements, which is indeed a concern — and already one with cell phones.
ITS and Spectrum
There were sessions on ITS and DSRC (the spectrum allocated for communications from vehicles to other vehicles and to local infrastructure points.) I’ve written on this topic before, and while I believe that of course robocars will make use of any such signals, they can’t be built to depend on them, and will in fact only encounter them occasionally for the next decade at least. As such, robocar development proceeds without much connection to the connected vehicle world, which I am sure bemuses those in the connected vehicle world.
A new robocar project named “Quasper” has emergence in France from the IRSEEM Esigelec lab and IFSTTAR. This vehicle uses a commercial actuator robot to control the wheel and pedals for drive-by-wire, and features a variety of typical sensors, though it only has a couple of smaller SICK LIDARSs rather than a high resolution LIDAR like the Velodyne used by many other projects. Their work is fairly basic for now. Because the DBW system occupies the driver’s seat, manual control is done through a joystick which controls the drive-by-wire robot. As a nice aesthetic touch, they have put much of the electronics in a standard roof box on top of their minivan. (Some of the early robocars did have a certain mad-science look with all their gear covering the roof.)
At the Consumer Electronics Show in Las Vegas this week, which I attended, most of the car announcements were for information and entertainment systems in cars. The one positive light is that many car vendors now realize they no longer can own and control all the digital systems in a car, and are making tools to connect with the driver’s phone and tablet for various purposes. They are not quite ready to use the phone as their data connection, though so most mean you will be paying for 2 data connections.
Mercedes and Audi had displays relating to ADAS and autonomous driving. While there was no sign of the Mercedes 2013 S class that will feature an autopilot, they did show a rather strange demo of a future self-driving car where the windshield was now a giant augmented reality screen. There, using hand gestures, you could point and objects in the real world and learn more about them. Of course, if your car is self-driving, you could use any interface, not just a gestural one. Today, gestural interfaces are not very well developed and most people prefer touchscreens and keyboards.
Audi showed their lane-keeping system and crash avoidance tools. While they didn’t make a deal about it, reports say they are planning to have a traffic-jam autopilot like the Mercedes one, but up to 60kph instead of just 40kph. This means combining their lane-departure-correction system, which steers you back into the lane using the steering motor, and their automatic cruise control which follows the car in front. It’s now up to Audio and Mercedes to see who comes out first. It will be very interesting, as well, to see how the public responds to it.
Mercedes is downplaying the value of a self-driving car. In a keynote by M-B boss Dr. Dieter Zetsche, he said the following (as reported by Motor Trend):
“We don’t believe that autonomous driving is the ultimate goal. We still want the driver to be in charge, in control, be the pilot. But when you are stuck in a traffic jam, when you are driving five hours straight through Nebraska, it might not be the most fun (to drive). And at that time you can switch and say, ‘Take over, I’ll read a nice book, look at movies, or listen to music.’ So it’s your choice, you’ll always be in control, but you have the choice to hand (the controls) over to an autonomous car.”
High end brands like Mercedes are in an uncomfortable situation. For a long time, they have sold their product based on the excitement of driving. They are torn between that message and the value of the robocar. My own prediction is that initially, people will indeed be keen on cars which do the boring driving, but take them to places where the driving is more fun. While fun driving will not go away, I predict people will start feeling that more and more driving is a chore to leave to the robot, and less and less is a thrill that demands taking the wheel. Look at how readily people who move to New York or other non-driving cities adapt to not driving. But that’s not a message sporty car brands will want to say. At the same time, self-driving will be a desired luxury feature and so luxury cars will have to offer it. The problem is that high end brands tend to offer both sport and luxury, often in the same vehicles.
The video is very science-fictional, though they have built a concept to look at without the auto-driving. Amusingly, they do show something I have always thought would be a nice ironic demo — playing a car racing game while in a self-driving car. While we are some distance away form a car where the entire surface, inside and out, is a display, I do think we’ll see display panels on robocars to help them act as taxis. Those display panels will say who the taxi is for, and might even have your favourite bumper sticker slogan while you move inside. Inside displays will be useful for all the things you would expect — dashboard, work and entertainment.
Toyota is also showing a Prius with a system them call AVOS (Automatic Vehicle Operation System.) While this is said to be a longer-term self-driving systems, report suggest that what will be done at the motor show is back-seat rides demonstrating parking ability and pickup at the door, similar to Nissan’s Pivo and Stanford’s Junior 3, but with some added obstacle avoidance. I have not seen reports of rides as yet. The Prius itself is use more basic sensors than the Google car and other major robocars.
Nissan has announced a new version of their Pivo concept car. The Pivo 3 here’s a story with a video offers 4 wheel steering and automatic parking, including a claimed functionality for automated valet parking. In the AVP case, the car requires a special parking lot, though it is not said what changes are needed. A few years ago the Stanford team demonstrated Junior 3 which could valet park in a lot to which it had a map, and which had no civilian pedestrians.
The Pivo 3, it is reported, “will come and pick you up when you summon it.” Presumably this involves both the parking lot and the path to the door where you summon it containing the special infrastructure it needs, but this is not described. What’s also described is something fairly important — automatic charging, where the car takes itself to a charging station and hooks up.
They say they have no commercial plans for the car, but that they do expect to put such functionality into other cars around “2016 to 2017.”
With the Tokyo motor show about to start, expect new announcements from Japan in the days to come — for example Toyota has promised a self-driving Prius at the show, in a similar parking lot mode to the Pivo.
In contrast to the optimism I usually present here, and last week’s article about a self-driving Mercedes just a year away it’s worth noting this interview with various BMW folks where they provide a much more cautious timeline of at least a decade. Part of their concern comes from the use of computer vision systems. These are much cheaper than laser scanners but do not provide the reliability needed; it’s no accident that all the successful teams in the Darpa urban challenge relied very heavily on laser scanning.
I’m enough of an optimist that I am ready to bring forward the question “When will a child be born that never drives because of robocars?” Of course there are many people in the developed world who never get a licence for a variety of reasons, particularly people who live their lives in Manhattan and other transit-heavy cities. But for most of us, getting a licence and getting on the road is a rite of passage. Yet studies are showing that teens are now waiting longer to get a licence with various reasons speculated.
Nonetheless eventually we will see somebody who would normally have been jumping at the chance to get a licence and get out on the road who never gets one because they have a robocar. It won’t be easy of course, since even those who have robocars will still need to travel to places that don’t have them and rent cars, but many people who don’t have licences today just make use of taxis and transit in those situations.
I will put forward the proposal that this child may already have been born. When I see a baby today, I wonder, “will this child ever learn to drive?” While 16 years is aggressive for the ubiquitous fully autonomous operation needed for this, I do think we’re on the cusp, and if that child has not yet been born, it’s not too far away.
One reason for this is all the forces that are already reducing teen driving. A teen debating whether to take the effort to learn to drive might easily be swayed not to because mom has bought him a robocar. Once a successful safety record for robocars is demonstrated, parents will buy them for teens — instead of buying them driving lessons — and pressure the teens to not take the risk of driving themselves.
In other news, here’s a pointer to work by designer Charles Rattray on the look of future robocars. His designs match with my position that many robocars should be half the width of today’s cars, carrying only 1-2 people, since the vast majority of cars today only carry 1-2 people. Today’s car buyers insist on 5 passenger sedans (or larger) but when you have mobility-on-demand you can use the right vehicle for the trip on every trip, and that’s going to mostly be one person vehicles. This in turn, is the real key to efficient transportation, because while you can do great things with more efficient or electric power trains and more aerodynamic cars, nothing compares to making the car smaller, lighter and narrower in a major way. He has many design sketches and a video of how he sees the cars in action.
For the first time, a car company has put a date on shipment of a car with self-driving ability.
According to British site Auto Express, Mercedes has revealed that their 2013 S-class will feature self-driving. Not clear if there is an official company press release, though the company has been talking about such features, as have many other companies. Realize that the 2013 model year is just a year away.
The car will feature radar based automatic cruise control, combined with lane-marker following, and the automatic driving will only operate below 40kph. In other words, this is designed to let you take your hands off the wheel in stop-and-go traffic jams, not to drive you at actual open driving speeds. You’ll need to pay attention to the road, not read a book, but at that low speed you’ll have decent warning if something goes wrong and the car starts drifting, so I suspect that in spite of warnings not to do so, people will get away with minor tasks like reading a few e-mails or even sending some.
While a very basic level introduction, this is still a milestone and will pave the way (love those road metaphors) for other companies. While the focus of the DARPA grand challenges and most visions of the robocar future has been on cars that can drive completely on their own, there are now strong signals that the technology will arrive in the form of driving assist, and human drivers will be called upon to still do much of the driving, in particular the tricky bits the systems aren’t safe to handle. In my article a few years ago the roadmap to robocars I suspected we might see a few specialized applications first, such as robot valet parking and even autonomous vehicles for military delivery applications, but now the autopilot is on track for showing up commercially first.
I’m just back from the “ITS World Congress” an annual meeting of people working on “Intelligent Transportation Systems” which means all sorts of applications of computers and networking to transportation, particularly cars. A whole bunch of stuff gets covered there, including traffic monitoring and management, toll collection, transit operations etc. but what’s of interest to robocar enthusiasts is what goes into cars and streets. People started networking cars with systems like OnStar, now known in the generic sense as “telematics” but things have grown since then.
The big effort involves putting digital radios into cars. The radio system, known by names like 802.11p, WAVE and DSRC involves an 802.11 derived protocol in a new dedicated band at 5.9ghz. The goal is a protocol suitable for safety applications, with super-fast connections and reliable data. Once the radios in the car, the car will be able to use it to talk to other cars (known as V2V) or to infrastructure facilities such as traffic lights (known as V2I.) The initial planned figured that the V2I services would give you internet in your car, but the reality is that 4G cellular networks have taken over that part of the value chain.
Coming up with value for V2V is a tricky proposition. Since you can only talk to cars very close to you, it’s not a reliable way to talk with any particular car. Relaying through the wide area network is best for that unless you need lots of bandwidth or really low latency. There’s not much that needs lots of bandwidth, but safety applications do demand both low latency and a robust system that doesn’t depend on infrastructure.
The current approach to safety applications is to have equipped cars transmit status information. Formerly called a “here I am” this is a broadcast of location, direction, speed and signals like brake lights, turn signals etc. If somebody else’s car is transmitting that, your car can detect their presence, even if you can’t see them. This lets your car detect and warn about things like:
The car 2 or 3 in front of you, hidden by the truck in front of you, that has hit the brakes or stalled
People in your blind spot, or who are coming up on you really fast when your’re about to change lanes
Hidden cars coming up when you want to turn left, or want to pass on a rural highway
Cars about to run red lights or blow stop signs at an intersection you’re about to go through
Privacy is a big issue. The boxes change their ID every minute so you can’t track a car over a long distance unless you can follow it over every segment, but is that enough? They say a law is needed so the police don’t use the speed broadcast to ticket you, but will it stay that way?
It turns out that intersection collisions are a large fraction of crashes, so there’s a big win there, if you can do it. The problem is one of critical mass. Installed in just a few cars, such a system is extremely unlikely to provide aid. For things like blindspot detection, existing systems that use cameras or radars are far better because they see all cars, not just those with radios. Even with 10% penetration, there’s only a 1% chance any given collision could be prevented with the system, though it’s a 10% chance for the people who seek out the system. (Sadly, those who seek out fancy safety systems are probably less likely to be the ones blowing through red lights, and indeed another feature of the system — getting data from traffic lights — already can do a lot to stop an equipped car from going through a red light by mistake.) read more »
Since getting involved with Google’s self-driving-car team, I’ve had to keep silent about its internals, but for those who are interested in the project, a recent presentation at the intelligent robotics conference in San Francisco is now up on youtube. The talk is by Sebastian Thrun (overall project leader) and Chris Urmson, lead developer. Sebastian led the Stanley and Junior teams in the Darpa Grand Challenge and Chris led CMU teams, including BOSS which won the urban challenge.
The talk begins in part one with the story of the grand challenges. If you read this blog you probably know most of that story.
Part two (above) shows video that’s been seen before in Sebastian’s TED talk and my own talks, and maps of some of the routes the car has driven. Then you get Chris showing some hard technical details about mapping and sensors.
Part three shows the never before revealed story of a different project called “Caddy”: self-driving, self-delivering golf carts for use in campus transportation. The golf carts are an example of what I’ve dubbed a WhistleCar — a car that delivers itself and then you drive it in any complex situations.
If you want to see what’s inside the project, these videos are a must-watch, particularly part 2 (embedded above) and the start of part 3.
There’s lots of other robocar news after the Intelligent Transportation Systems conference, which I attended this week in Orlando FL. The ITS community is paying only minimal attention to robocars, which is an error on their part, but a lot of the technology there will eventually affect how robocars develop — though a surprising amount of it will become obsolete because it focuses on the problems caused by lots of human driving.
The list of robocar teams grows again with a new project from Oxford university, led by Paul Newman. Nissan is also involved, though the base vehicle is a Bowler Wildcat off-road vehicle.
The project sports a LIDAR design I have not yet seen, with 4 laser units on a mount spinning at what looks like 1-2hz, but they claim a 40hz sampling rate and do have very nice mapping results. They claim their localizer is very good, and demos show it working on rough off-road terrain. Some videos also see it doing waypoint driving without the LIDAR but they talk about why GPS is not adequate.
The claims about the vehicle have a British understatement. They say it will be 10-15 years before it’s ready for the roads, and talk mostly about simple problems like handling traffic jams — something Audi, BWM and VW have all claimed they will release in the middle of this decade, using simpler sensor systems. He also envisions a future arms-race where a car that can do 10 minutes/day of self-driving competes with one that can do 15.
Congestion is their main message it seems, citing the Dept. for Transport’s figures of a 25 billion pound cost for congestion in 2025 in the UK.
Boston Dynamics has gone even further with their latest model, AlphaDog
The AlphaDog’s legs are hydraulic, and so adding legs like this to a car which has a motor and compressor is not so far fetched. In this design they could easily fold up into the sides of a single person wheeled vehicle. In the video, the robot is shown carrying 400lbs of weights, and a range of 20km is claimed. You might not quite want to ride it yet, but that’s coming.
Let’s look at some of the consequences for transportation and cities:
Houses need not be on streets to have full access by small vehicles and cargo delivery robots. They can be on the side of hills and up stairs. Neighbourhoods can be built with just small lightly paved or graded paths so that the robot’s legs don’t disturb the terrain.
The robots may well, in a controlled environment, be able to place their feet with good precision. As such the path for a walking robot might look like just a series of stone pads dotting the grass — the way some paths for people look. In reality they would be more sturdy, but that’s what they could look like.
In developing countries which do not have infrastructure, they may never have to put in that much infrastructure. Combined with flying robots, delivery of goods can become possible to any location, and at high speed.
The world’s tourist destinations may become swamped with people who can ride a walking robot to remote locations where before the daunting hike kept the crowds down. There will be efforts to ban walking chairs, but the elderly and disabled will be able to fight such bans as discriminatory.
Indeed, for the disabled and aged, the walking chair robot might well open up lots of the world that is now closed. The main issue would be power and noise. The motors that power BigDog are very noisy. AlphaDog in the video is using external power.
Robotic cargo delivery (deliverbots) need no longer be limited to places you can roll up to. That can include places inside buildings, even up stairs.