The flurry of predictions about the rosy future for self-driving cars has missed an important feature: the car needs to be told the driver’s destination before starting out, just as pilots of airplanes have had to do pretty much from the beginning. How will this change ground transportation? And how soon with drivers (or their self-driving cars) be required to ‘file’ this plan with the local traffic planners? Airlines do that now with a federal agency. My experience with carsharing showed how even the scant commitment users had to make — for when they would leave and when they would return (and the $25 fine for returning the car late) — had the effect of injecting rigor into trip planning.
The drivers of most buses and larger trucks already do that, even though there is no government agency requiring it. That is because their vehicles are owned by — or are contracted to — large concerns which have made commitments to third parties about what/who will be delivered where by when. It’s called “logistics” and is now a science. Even though the federal aviation authorities use the mass of schedule information to “manage traffic” in the air, this is not pooled by ground-transportation authorities to allow for congestion-avoidance strategies, leading to, at least, greater efficiency.
If each vehicle — self-driving or human-piloted — sent their users’ plans to a central website that could do on-the-fly projections of the amount of traffic at each pre-identified “pinch-point” in its system, some very beneficial steps could be taken:
– fewer driver last-second maneuvers would occur, some of them causing collisions or near-collisions; and no drivers would get lost;
– traffic planners could ‘see’ congestion before it would occur, allowing time to advise drivers or on-board computers to alter their plans according to scientific suggested alternative routes.
– drivers with empty seats — sensors in the cars to detected the status of each seat, as they now can sense whether a seatbelt should be deployed — could function as a provider of “trans-seat” for people on foot needing a time-saving lift as part of their own trip, a kind of “Uber-ization” of transit — or the automation of hitchhiking to make it work on an intra-urban basis — all linked by our smart-phone system.
Already, proponents of self-driving cars are suggesting that they will significantly reduce road collisions and car-ownership (thus a city’s car ‘population’) by eliminating both driver error and vehicle down-time (it would drive itself from one user’s trip-destination to next user’s trip-origin) eliminating the between-trips parking requirement associated with private car-ownership. The plans of users will be known before a shared-smartcar company even assigns the trip, further improving vehicle efficiencies.
Would we, today, consider travel by air, train, or bus if such a system didn’t exist for each? We even insist that it is not enough to now a vehicle’s schedule, but its current status, too (since plans can be adhered to flawlessly). And if the road system planners, knowing the plans of all vehicles on the road and those about to enter the roads, are able to predict congestion, and share its findings with all users, wouldn’t that allow people to make — and change — their plans as traffic status report became available, including delaying or dropping their upcoming plans to drive, at least for the next few hours? Might, finally, we shape travel demand to road capacity, rather than expecting government to eternally expand it to our whims-of-travel?