I read this rather breathless article about driverless cars a couple of days ago. I was smitten with the focus on parking. Less written about, at least here, is that much of the advantage in driverless cars is in much tighter following distance. The driverless cars that are out there now, driven over 400k miles in California’s countryside and cities, are lined with sensors to detect and avoid collisions. The dream of driverless cars is that all cars are driverless, in the same way that all cars drive on the right today. At that point, they won’t need sensors to avoid collisions with each other, as they will be able to use transponders for that. I assume the subroutine will be that transponders will allow much closer following, but that sensed objects off the network will be given a wide berth. They will still need the sensors to avoid collisions with the uncouth and unlinked masses, like old cars, deer and walkers.
The real problem with driverless cars is that they are still cars. They still get about 30-50 miles per gallon of gasoline. They still perform best in many-to-many networks, meaning average vehicle occupancy will never be that high. They still require about 1% of an acre per passenger. Being driverless, their average speed might be well above 35 MPH. Great for passengers, terrible for anything that gets hit by them. They still make us dependent on a mode that spaces out and abstracts land use, rather than concentrating and intensifying it. They still make us dependent on enough energy to move 300 pounds of people in 3,000 pounds of vehicle. They still make us pay rent for the privilege of being on the network, and make those not on the network second class citizens. The cost of that requirement will probably be over $10,000 every year by the time driverless cars really get going. We could use that energy for computing or buildings, or growing food. But no, we have to push the heavy vehicles all around the cities.
So what are driverless cars good for? They would likely be more fuel efficient, from better management of stop and go traffic. They would be more time efficient in route and in parking. The linked article harps on that. They would almost certainly be safer than human piloted cars, even for walkers and bikers. They could also make cars function more like services, not possessions. A car club could rent its members small, efficient, single or two-seater cars for most trips, and offer them a truck or SUV only when they needed to haul larger groups or cargo. There would be no need to worry about skill, and people wouldn’t be driving large trucks for routine trips on the contingency that they would need that size truck every month or so**.
As long as they keep all those sensors around. And of course they would be yet another luxury in our 20,000 year tale of technological progress. I see little reason why driverless cars should not be the majority of traffic in under 20 years.
** Excellent suggestion via Lars Fields.