The news articles about self-driving cars are coming relatively fast and furious considering that the closest thing to autonomous cars commercially available on the road today are cars that can parallel park themselves. This April 10 article, Former General Motors R&D chief says self-driving cars will be on sale by 2020, points out why, considering that it suggests that the plans are to go from the current self-parking cars to close to fully autonomous cars in just 8 years. In an excerpt from the article, Larry Burns, the former GM R&D chief, states that he “sees a day when true driverless cars with no one on board will be shuttling about servicing many passengers over the course of a day, rather than sitting around for hours parked and waiting for their owner.” I’m not sure if the current powers at GM are looking forward to that vision, as it seems to point away from the concept that every citizen should have their own personal car. I am certain I am not the only individual realizing that would mean much fewer auto sales.
The plans apparently are for a slow transition toward fully autonomous, with Self-driving Cadillacs coming mid-decade. Considering that we are already two years into this decade that sounds unrealistically soon to me, even for the less than fully autonomous “super cruise-control” that is proposed in the above article. However, the technology to implement self-driving cars exists, and it seems to me that the biggest obstacle will be public acceptance.
There was no indication in the airlcte how much driverless systems would cost. Equipping cars with all those lasers, sensors, GPS navigation systems, artificial intelligence and who knows what else will be expensive.=================================================In the end, it will be much simpler than this. Much of those sensors are to figure out where the road is, and then to figure out where the other cars are. This is called non-cooperative control.Imagine a bunch of RFIDs embedded in the road, like those little reflectors, each one telling the car unambiguously, you are here, you are here, you are here, intersection ahead. A lot of really hard sensor technology goes away.Likewise, you need not sense your eight nearest neighbors if you are talking to them: Iam slowing down, I am speeding up, I am turnng left, I am changing lanes. This is the sway the new aircraft control system will work, with collision avoidance worked out cooperatively and as needed, diverting each aircraft th minimum amount necessary.At one swoop, rear end collisions will be eliminated. The uncertainty associated with evaluating the next drivers behavior will be gone. What would be a breathtaking merge for three drivers will be commonplace with negotiated cooperation. This might easily increse the throughput of the roads by a third, while completely eliminating congestion, which is mainly an issue of uncertainty. In times of high volume traffic may slow down but the travel will be smooth and steady, and certainly faster than now. Area wide rerouting may take place with the help of satellite or drone based sensors.What we are seeing is a feasibility stage, the final technology will be much easier and cheaper.Rear end collision avoidance is already here, but again, it is noncooperative.A cooperative system will be much faster and much cheaper.
Thank you for the highly informative and knowledgeable reply. Please note that while I plan to leave this site and blog up, I have converted it to a WordPress site at http://www.charles-koeppen.com. I am mirroring the blog posts between the sites, and since your comment is so highly informative I will re-post it there with full accreditation.