Tag Archives: driverless cars

Driverless buses could improve mass transit

Posted on by

Discussion of driverless buses in Britain highlights the efficiency they could offer, leading to improved service:

Claire Perry, the Transport Minister, said that operating buses without drivers could help companies provide “better and more frequent” services, particularly in rural areas.

She also revealed that work is already under way to identify any problematic “regulatory issues” which could prevent the vehicles being rolled out on roads across Britain.

Speaking at the Driverless Vehicles Conference at Thatcham on Wednesday, Mrs Perry said she could “see a future where driverless buses provide better and more frequent services”.

“A major component of rural transport is the cost of the driver – and so a truly driverless bus could transform rural public transport in the future,” she said.

Driverless cars offer safety and commuting convenience but this is a twist: mass transit could be more frequent and cheaper without drivers. It would be interesting to know how much cheaper this could be. Would this mean a 20% increase in bus service for the same price or is it something even more drastic? If so, perhaps this could make buses a lot more attractive, particularly in rural or suburban areas where riders may not necessarily want to ride with a lot of other people and want service that doesn’t inconvenience them much.

California issues first autonomous car driving permits

Posted on by

Several corporations recently received California’s first autonomous car driving permits:

On Tuesday, Audi became the first car manufacturer to receive a California autonomous car driving permit (as of this writing, Mercedes-Benz and Google have also filed for and received permits). The permit was presented to Audi by Sen. Alex Padilla, who signed the state’s new autonomous vehicle laws that went into effect Tuesday; the law will allow for the legal testing of autonomous vehicles on public roads…

One is the specific mention of a visual indicator that clearly signals to the driver when autonomous mode is engaged. Making sure the driver is completely familiar with the technology and understands when the car is under machine control versus human control is something carmakers must get absolutely right. Consider what GM is doing with its Super Cruise technology, which allows the car to take over steering and pedal operations in certain highway conditions. Earlier this month GM announced that Super Cruise will be available in select 2017 model year cars. Those cars will likely have the same indicator that we experienced when testing Super Cruise—a large light bar on the top of the steering wheel that indicates when the car is in control (green), when the driver needs to take over (red), and when the driver has control (blue). Hard to miss that. Oh, and it issues an audible alert as well.

Something else to consider: According to the permit, should the driver be unable to take control of the vehicle during an emergency or system failure while autonomous mode is engaged, “the autonomous vehicle shall be capable of coming to a complete stop.” Pretty important! But also a little scary when you think about a car just stopping on the highway. After all, the permit doesn’t say the car must be able to safely pull off the road and come to a complete stop. And in reality, that’s probably asking a lot for now. It’s a reminder that if we want to test autonomous vehicles in the public domain seriously, we have to understand there will be risks…

Lastly, it’s worth noting that the permit calls for an extra device—separate from the data recorders already required in cars—to specifically monitor and record the autonomous systems and their sensors. On top of that, the information must remain accessible for three years. As optimistic as lawmakers and auto manufacturers are about the potential for autonomous vehicles, they also know that one bad accident could stymie progress and reaffirm the public’s worst fears. In case an accident does happen—and eventually, it will—at least they’ll know exactly what went wrong.

Some interesting extra pieces to these permits. All of this suggests that there are still some important things to sort out before driverless cars hit the roads in large numbers.

A few other possible additions that came to mind:

1. An indicator on top of the car or with the front and back lights that shows other drivers that the car is in autonomous mode. We haven’t heard much how such vehicles would change their behavior based on the drivers around them. Say someone doesn’t like their speed and so they tailgate the car, an action that sometimes leads to the front driver speeding up. What would an autonomous car do?

2. A running set of easy-to-understand output from the autonomous car to the driver. It is one thing to provide an indicator that the car is running itself but another to give feedback to the potential driver. Granted, these vehicles are likely making a ridiculous number of calculations per second but I’m guessing some users would like to know what the car is “thinking” as it acts.

Could a 220 mile round trip commute become more common?

Posted on by

Supercommuters are rare but here is one man’s story of driving 220 miles a day to a “dream job at the Department of Justice.” Of course, he says it is all for his family.

My first thought: this has to have long-term consequences for this man and his family. Those miles and hours will add up.

My second thought: I wonder if these sorts of stories will pick up with the rise of driverless cars. Right now, the problems of a long time spent in the car include lost time and stress. But, imagine you get in the car in your driveway, point to the destination, you don’t have to touch anything, and you get to relax and do what you want in the cabin until you arrive. Perhaps the driverless car will even lead to an uptick in driving, reversing a trend that threatens gas tax revenues. For those who like driverless cars for their gains in safety, would they also be willing to accept more driving?

Experts: cars in the near future won’t have mirrors, pedals, steering wheels…

Posted on by

A survey of over 200 experts suggests driverless cars of the next few decades will be missing some now-common features:

By 2030, most new cars will be made without rearview mirrors, horns, or emergency brakes. By 2035, they won’t have steering wheels or acceleration and brake pedals. They won’t need any of these things because they will be driving themselves…

The shift to cars without steering wheels and pedals will be revolutionary. It’s one thing to get a driver to let go of the wheel on long highway drives or a boring commute. It’s quite another to put him in a car that he can never drive, even if he wants to.

The change is inevitable, says Alberto Broggi, a professor of computing engineering at the University of Parma and an IEEE fellow. Cars that don’t need human drivers anymore will shed parts made for human control. “There’s nothing you can do about that.” The change will free auto design from the rules that have constrained it for a century. (Only Google has publicly addressed the idea, with a prototype it plans to start testing on public roads this fall.)

This all makes sense if the cars drive themselves but it could be quite a change. Will it really free up designers to create something different than what we have now or will the basic shape remain the same with an altered interior? There is a lot of potential here to create something that doesn’t look like a car as we know it.

How Google’s driverless car navigates city streets, construction, and urban traffic

Posted on by

Eric Jaffe provides some info on how driverless cars navigate more complex urban roads:

Boiled down, the Google car goes through six steps to make each decision on the road. The first is to locate itself — broadly in the world via GPS, and more precisely on the street via special maps embedded with detailed data on lane width, traffic light formation, crosswalks, lane curvature, and so on. Urmson says the value of maps is one of the key insights that emerged from the DARPA challenges. They give the car a baseline expectation of its environment; they’re the difference between the car opening its eyes in a completely new place and having some prior idea what’s going on around it.Next the car collects sensor data from its radar, lasers, and cameras. That helps track all the moving parts of a city no map can know about ahead of time. The third step is to classify this information as actual objects that might have an impact on the car’s route — other cars, pedestrians, cyclists, etc. — and to estimate their size, speed, and trajectory. That information then enters a probabilistic prediction model that considers what these objects have been doing and estimates what they will do next. For step five, the car weighs those predictions against its own speed and trajectory and plans its next move.

That leads to the sixth and final step: turning the wheel this much (if at all), and braking or accelerating this much (if at all). It’s the entirety of human progress distilled to two actions…

The Google car is programmed to be the prototype defensive driver on city streets. It won’t go above the speed limit and avoids driving in a blind spot if possible. It gives a wide berth to trucks and construction zones by shifting in its lane, a process called “nudging.” It’s extremely cautious crossing double yellows and won’t cross railroad tracks until the car ahead clears them. It hesitates for a moment after a light turns green, because studies have shown that red-light runners tend to strike just after the signal changes. It turns very slowly in general, accounting for everything in the area, and won’t turn right on red at all — at least for now. Many of the car’s capabilities remain locked in test mode before they’re brought out live.

Quite a process to account for all of the potential variables including other drivers, pedestrians and cyclists, weather conditions, and other objects on the road like construction or double-parked vehicles. I imagine this is some intense code that has to provide a lot of flexibility.

This also reminds me of some of my early experiences driving. It took some time to adapt to everything – watch your speed, check all those mirrors, what are the other cars doing, what is coming up ahead – and I remember wondering how people could even carry on conversations with others in the car while trying to drive. But, with practice and adaptation, driving today seems like second nature. And, I suspect from my own experience that drivers are not 100% vigilant (maybe 80% is more accurate?) while driving as they generally think they have things under control.

All that said, driving is a remarkable cognitive task and replicating this and improving on it in a 100% vigilant system requires lots of work.

Uncertainty over who is liable for crashes of driverless cars

Posted on by

When an autonomous car gets into a crash, who is responsible? This question may just delay their mass market release:

“Automaker liability is likely to increase. Crashes are much more likely to be viewed as the fault of the car and the manufacturer,” Anderson said. “If you’re an automaker and you know you’re going to be sued [more frequently], you’re going to have reservations.… The legal liability test doesn’t take into account the long-run benefits.”

In other words, even though a technology is an overall boon to the greater good, its rare instances of failure—and subsequent lawsuits—won’t take that into account. That could slow the movement of driverless cars to the mass market if automakers are wary of legal battles…

As they grapple with what autonomous vehicles might mean for their industry, the legal frontier remains uncertain as well. One possible solution? A payout fund set up to compensate victims of driverless car accidents. That could be modeled similar to the Health and Human Services Department’s vaccine injury compensation fund, which takes a 75-cent tax from every purchased vaccine. The no-fault program helps those who have been hurt by vaccine-related incidents without exposing the medical community to legal battles and expensive damages payouts.

In the early stages, subsidies may be required to help driverless cars take hold in the market, according to Rand’s report on the technology’s adoption. Part of the money allotted for that could be set aside to help potential victims.

Sounds like there is still some work to do here and automakers are quite aware of these issues with recent events like the $1 billion settlement payout from Toyota. While it sounds like the technology is getting close, the legal and social issues might also prove difficult to nail down. But, the outstanding safety potential of driverless cars may force a quick resolution to the liability issue in order to save lives sooner.

Smart highway features coming to two Chicago highways

Posted on by

A stretch of I-90 will be a “smart highway” within several years:

By 2016, the Tollway plans to install an elaborate system of sensors, cameras and overhead signs on a heavily traveled stretch of the Jane Addams Memorial Tollway (Interstate 90) between the Kennedy Expressway and Barrington Road in Hoffman Estates.

The plan is similar to, but more sophisticated than, a $45 million initiative that the Illinois Department of Transportation will implement during the next two years along the Edens Expressway and the northern stretch of U.S. Highway 41.

The Tollway plan includes installing signs with red and green signals over each lane at every half-mile that would advise motorists about safe speeds and warn of lane closings from accidents or breakdowns…

The goal is to make the Addams, which handles about 317,000 vehicles a day, “a true 21st-century, state-of-the-art corridor,” Tollway officials say.

The fiber optics and other infrastructure being installed on the soon-to-be rebuilt stretch of tollway will be able to accommodate even more sophisticated technology, which might someday automatically drive cars, officials say…

Tollway officials said Washington state’s experience with ATM has been compelling. The system is in use on I-5 in Seattle and on I-90 and State Route 520 between Seattle and Bellevue, and since 2010, the Seattle area has seen an 11 percent decrease in primary accidents and a 40 to 50 percent decrease in secondary accidents, officials said.

While highways in the United States are an engineering marvel, the lack of information about conditions on them has always struck me as a bit odd. It sounds like this new system is intended to provide information for two main purposes: warn people of upcoming obstacles which could then lead to fewer accidents and also to tell people of slower travel times so they can then make decisions about what roads to use.

Up to this point, motorists have been limited to varying levels of information:

1. You see what is front of you. Sometimes, you can spot some of these problems a long way away and get off sooner. But, too often, the line of sight is blocked and before you know it you are in a slow stretch without any alternatives.

2. Traffic reports on the radio. The veracity of these reports can vary.

3. Traffic data now available on GPS and smartphones. These seem to be generally accurate.

4. Cameras along heavily traveled routes. For example, see this set of images from cameras along I-80/94 at the bottom of Lake Michigan. This is more useful these days with smartphones.

Of course, this article also hints that this may just help set up the infrastructure to have completely smart cars where all of the information may be wirelessly passed between cars and limit the human dimension all together.

Good question: “What Will Happen to Public Transit in a World Full of Autonomous Cars?”

Posted on by

The fate of mass transit is unclear in a world of all autonomous cars:

The question of what they’ll mean for transit was actually on the program this year at the Transportation Research Board annual meeting in Washington, where several thousand transportation officials and researchers met to talk about state-of-the-art asphalts, biker behavior, and the infrastructure of the future. In one packed session, I heard Jerome Lutin, a retired longtime New Jersey Transit planner, say something that sounded almost like blasphemy.

“We’re just wringing our hands, and we’re going to object to this,” he warned the room. “But the transit industry needs to promote shared-use autonomous cars as a replacement for transit on many bus routes and for service to persons with disabilities.”…

The implication in this raises (at least) two more questions: Exactly where (and when) will it make sense for people to use buses or rail instead of autonomous cars? And if autonomous cars come to supplement these services, should transit agencies get into the business of operating them? In my initial daydream – where shared self-driving cars are whisking us all about – it’s unclear exactly who owns and manages them.

Lutin sounds skeptical that transit agencies will be able to move into this space. “They don’t adapt well to change,” he says. They’re also governed by rigid mandates that limit what they can do. A mass transit agency can’t overnight start operating something that looks like a taxi service. Public agencies also must contend with labor unions, and labor unions likely won’t like the idea of replacing bus routes with autonomous cars.

This does seem to trade a public good – mass transit paid for by taxpayers and users – for private goods, autonomous cars owned by individual users. While we haven’t seen prices for driverless cars yet, I can’t imagine they are going to be too cheap at the beginning. Even a less appointed driverless car, say a Chevrolet Aveo, is going to need more complicated gadgetry to be autonomy. But, as this planner notes, Americans do tend to like more private transit options if they can afford it.

Driverless cars might lead to the safest roads ever seen and highways that actually work

Posted on by

Google’s Sebastian Thrun discusses the safety advantages of driverless cars as well as effective highways in this video. Two quick thoughts:

1. Autonomous vehicles can help stop the wave or accordion pattern of driving where someone slowing down at the front of a line of cars backs everyone up. Thrun talks about a a much more closely coordinated zipper kind of merging where cars going basically bumper to bumper can accept new vehicles with little change.

2. Even with a dip in recent years, more than 30,000 Americans die each year in vehicles. Strun says most of this is due to driver error, which can basically be eliminated when humans no longer control the car.

All of this sounds good…

Study: self-driving cars would be expensive but would save lots of lives, money down the road

Posted on by

A new study suggests self-driving cars would have a lot of benefits:

If 90 percent of vehicles were self-driving, as many as 21,700 lives per year could be saved, and economic and other benefits could reach a staggering $447 billion, said the study, a copy of which was provided to The Associated Press…

For example, the passenger compartment may be transformed as former drivers safely work on laptops, eat meals, read books, watch movies and call friends. And cars that can be programmed to pick up people, drive them to their destination and then park by themselves may change the lives of the elderly and disabled by providing critical mobility.

Once a critical mass of self-driving cars is on the road, they can start “platooning” — driving closely together but keeping a steady distance between each other without the fuel-burning, time-wasting, stop-and-go typical of traffic congestion. That could smooth traffic flows, reduce commute times and increase highway capacity.

Government research indicates driver error is likely the main reason behind over 90 percent of all crashes. Over 40 percent of fatal traffic crashes involve alcohol, distraction, drugs or fatigue. But self-driven vehicles wouldn’t fall prey to such human failings, suggesting the potential for at least a 40 percent reduction in fatal crashes, the study said.

Who doesn’t like safety, saving money, and reducing the stress of driving? If more companies produce such cars and they become legal as well as widely accepted, it will be interesting to see how quickly prices might drop.

I wonder at the efficacy of self-driving cars if they only constitute a small number of cars. While they would still allow drivers to do other things in the car, they wouldn’t quite be as safe until a majority of cars are self-driving. It seems like the benefits rise at a steep curve: a few self-driving cars won’t help much but having 90% self-driving cars would be transformational.

I also am curious to know whether there would be people who would oppose self-driving cars because they can perpetuate car culture in the United States. If peak oil is not near and self-driving cars make driving safer and easier, might Americans have cars for centuries?