Robocar legality in Nevada, Germany and more news

Topic: 

A wrapup of robocar news from the past couple of weeks:

Nevada governor Brian Sandoval rides in Google Car

After Nevada's recent legislation directing their DOT to explore legal operations for robocars in the state, the governor "took the wheel" of a Google car. Very positive impressions from the governor and DMV head.

Video of Virginia Tech "car for the blind"

Here's a video interview with Dennis Hong of Virginia Tech on his effort to make a car for use by the blind.

Autonomous car team in India

A new student robocar team has sprung up in India. They're still early but their goal of driving in the crazy Indian traffic is a daunting one. Robocars have many advantages at low speed, where the 360 degree vision of LIDARS makes them see more than a human will. Harder is modeling the behaviour of other vehicles and playing games of chicken.

More mainstream press articles

Mainstream press articles of the robocar future and the intermediate technologies are growing in number. Here's Smartmoney on near-term technologies and a Slate piece that, like almost all mainstream press pieces, asks whether people are really willing to give up the freedom of driving. Perhaps I'm too immersed, but in my immersed perspective I have simply stopped wondering about this. There will be a few who think like the dodge ad but huge numbers of people keep asking me when they can get one.

IEEE conference at Stanford paints alternating views with optimism vs. long roadmaps

Last Saturday a small IEEE conference at Stanford covered car automation technologies, including a morning on autonomous vehicles with mixed views. Steven Shladover, for example has a decades long history in important projects like cars guided by embedded road magnets, ITS, cooperative cruise control and platooning, but he is highly skeptical of autonomous cars which drive with regular cars, insisting instead that dedicated lanes are the answer. He believes this will start by building dedicated lanes for express buses (BRT) -- which is something there is political will to do in many cities -- and then automating the buses in those lanes. Once this is done, cars can enter the lanes if they communicate properly with the other vehicles in the lane and the lane itself.

This infrastructure approach is simpler from a technical standpoint, but the building of new infrastructure is such a hard problem and point of slow progress that my bet, as readers know, is on robocars on ordinary streets. Without the BRT component, I view proposals for new robot-only lanes to be dead in the water. Still, it's worth paying attention when somebody with lots of experience disagrees so fundamentally with your views.

Volkswagen, while having recently promoted their Temporary Auto Pilot, displayed a roadmap that was much slower, suggesting that having a car that could pick you up at the airport or park itself on streets was something we might see in 2028.

Another lesson from the conference was the extreme difficulty of introducing radical innovation through big automakers. Cars are perhaps the most complex product sold, as well as the most expensive consumer product for most. As a result the industry has created huge amounts of "process" to how it plans and innovates, and that process is not ready to accept much in the way of disruptive technology. As I wrote earlier about the radio as the potential place for innovation in cars, car makers are now considering the central console where the radio and other controls are found the "golden stack" and they want to be the provider of it. Especially because the stuff they sell there sells for a huge margin; people often pay $2000 for an in-car GPS that's worse than what they get free in their phone or for $250 in the aftermarket.

German team gets permission for their robocar tests on city streets

The AutoNOMOS team at Freie Universität Berlin reports they have been approved to test on city streets. This testing will be similar to the testing Google has reported doing in California, with a safety driver and copilot in the car to monitor and take control in any situation that presents a safety risk. According to the New York Times, Google didn't seek a specific permission but state officials did agree, when asked by the times, with the interpretation that a vehicle with a licenced driver responsible for vehicle operations was legal.

Porsche trying to make a very smart cruise control

While not up to Volswagen's temporary auto pilot, which combines ACC with lane-following, Porsche is developing a learning automatic cruise control that will come to understand road curves and speed and drive better as it learns.

Lots of exciting news, even in the slow summer season. Disclaimer note: The Google car project is a consulting client of mine.

Comments

The Sub-title of this article, "Nevada governor Brian Sandoval rides in Google Car," made me think that the acceptance of robo-cars should come in the form of a Turing Test adapted for automobiles. It would work like this: construct a blind such that the front passenger cannot not determine if a human driver is present. The vehicle, human or non-human driver and passenger go for a ride on a set course. The passenger goes for multiple rides not knowing if the driver is human and must make that determination at the end of the trip. Perform this test over a sufficiently large passenger population, a variety of traffic conditions and use statistical analysis to tell whether robocars are ready for acceptance.

Alternatively, it could be called the Folger's Crystals Driver's Test. But I think T[o]uring Test has a nicer ring.

But success at this test is very close, and it's not really a good safety test. Humans have an accident every 250,000 miles or so so you will not be judging not safety but ride comfort. Before long I suspect the robots will be able to give very comfortable rides, and you could program one to drive as recklessly as a human in order to pass the test, but otherwise would not. Humans have bad driving habits and can't see 360 degrees. Robot flaws will come from unexpected situations, while human flaws come from inattention.

This technology is not catching on, because there is no demand. This is proved by the fact that the men advocating it are about 100 pounds overweight, like Brad Templeton is.

These men have a Jabba the Hutt fantasy of getting wheeled around on an automated platform. That is why they want this so badly, but normal people don't.

Talk about ad-hominem.

Totally unnecessary, and not at all accurate.
I ride my bike to work, and am still totally excited about robocars.

Partly because of safety, I love the idea of being able to improve software to reduce road accidents, while we can't easily reduce the road accidents for human drivers other than slowing them down.

Partly because of convenience. I love the idea of being able to take the car to the city, and then let it drive itself home, rather than needing to inconvenience a human giving me a lift.

Partly because I don't have enough free time in my life. When I'm driving my 1.5yo son around, I'd MUCH rather spend the time playing with him rather than watching the road. I value my time with him enough that a half hour trip would be significant to me.

Partly because of the new uses for cars that will appear once they're automated. Ambulance cars that could rush appropriate emergency supplies to first responders if there isn't a nearby hospital, home delivered food (If I could get *healthy* home-delivered food for a reasonable price I would eat healthier in general than when I cook for myself. And removing the cost of paying a driver will help make that achievable.) More efficient public transport (Why run a huge bus which carries three passengers. How much better for the customer and environment if I can order a bus on my iphone, and depending on the demand and destination it sends a small car or a large bus.) and so on. And these are only ideas that I came up with while writing this, the real future will probably be way more interesting.

I don't lament that I use a car instead of a horse to get places. I don't lament that I listen to my music through a speaker rather than needing a human orchestra. And I won't lament not needing to be constantly holding a steering wheel, and being responsible for my safety, my passengers safety, and the other commuters safety, while driving somewhere.

Plus, y'know, when I'm riding my bike to work, I bet robocars will be more courteous than human drivers, many of which seem to be blind when it comes to bikes.

Add new comment