Roads That Work for Self-Driving Cars
wsj.com
wsj.com
http://www.gizmag.com/volvo-road-magents-autonomous-cars/311...
It wouldn't be too hard to pattern the magnets to communicate other info to the car too, such as an upcoming sharp turn. More abstractly, magnets could mark the car's position extremely precisely, and the car could then acquire regularly updated info about upcoming road conditions from the internet, or a short-range local radio broadcast.
Magnets seem very cheap compared to the cost of resurfacing. I don't understand why we haven't been putting these in for years.
> The magnet-based system relies on magnets embedded along the center of the lane 1.2 meters apart to define the roadway. The car tracks the magnets with less than 7.5 centimeters of error. Although this requires doing something to the road, it is very cost-effective to spend about $10,000 per lane mile to buy and install the magnets compared to spending from $1 million to $100 million per mile to build new lanes. And the life of the magnets is 30 to 50 years -- longer than the best pavements.
http://www.fhwa.dot.gov/publications/publicroads/97novdec/p9...
If a car can work if the roads are well marked, run it can work in many different countries too. That's a huge boon for car makers.
Additionally, humans do fail to keep track of lane lines in bad weather, and this result in accidents that could be prevented if magnet-directed computers were driving the car.