My father can remember when cars shared the streets of his town with horses. Some people still had carriages and a good number of delivery wagons (ices, coal) were horse-drawn vehicles. In 1915, the year of his birth, there were over 26 million domestic horses and mules in the US. By 1960, that number was about 3 million.
Though the decades are different, this looks like the kind of radical change Neal Stephenson cited in his “big stuff” challenge. A collision in my head between my father’s recollection and an article in The Atlantic about Google’s self-driving cars prompted me to take a closer look.
Horses haven’t disappeared in my father’s lifetime, and I suspect human-driven cars won’t disappear in mine. But I know I’ll see human drivers share the road with self-driving cars, and I strongly suspect I’ll see the latter come to dominate the highways, if not the secondary roads.
The benefits are already being proclaimed. Fewer accidents. Higher density without traffic jams. Energy efficiency. People have also pointed out that a key source of employment for people without university degrees will disappear. (There are about 230,000 taxi and limo drivers in the US and about 3.5 million truck drivers.)
Clearly, the social and economic impact of this changeover in transportation would be dramatic. Would it qualify as “big stuff”? Let’s view it through that lens:
Neal Stephenson’s criteria for such projects are these:
- They should be achievable, without basic constraints (like time travel or faster than light travel).
- It’s okay, even desirable, that they should require the work of a career to complete.
- They should have a practical purpose.
- It’s okay for the project to be a technological fix to a major human problem.
The ideal. In the year 2054, human drivers are as common as people riding horses today. Overwhelmingly, driving cars is limited to entertainment and hobbies.
As a user of self-driving vehicles, my regular use – for shopping, meeting friends, commuting to work – will be programmed in. Even better, the arrival of vehicles to my doorstep (or a designated access point, perhaps so I can get some exercise walking) will be synchronized to my real availability, based on reports from my smart house on what my readiness to travel is. The vehicle itself could be a Segway-like two-wheeler or a truck, depending on my needs. If there are several stops, my travels will be optimized for time, energy consumption, the number of people who will be traveling with me, the schedules of people I’m meeting, and personal needs (such as lunch or a coffee break). I’ll also have the opportunity to choose an alternate route to, say, enjoy scenery.
How? None of this is very far beyond reach today. The cars need a more complete infrastructure (especially of information, but Google is working on that). What’s involved otherwise is mostly advanced logistics and sensing, with input from a network of devices. Optimization and scheduling for transportation is already close to what’s needed. The trickiest aspects may be legal constraints (with social opposition) and crisis response (when extreme weather, social unrest, and earthquakes force accommodation and rapid adaptation of systems).
The intermediate. In the year 2024, human-driven and self-driven vehicles will share highways. All the human-driven cars will have the capacity to preempt the drivers to avoid accidents and to inhibit choices that would cause problems to the system (such as driving recklessly). Those who abuse the driving privilege will be stopped from using highways (temporarily or permanently). Most secondary roads will still be dominated by human-driven vehicles. The “last mile” for trucks and taxis will be taken over by humans who will man to vehicles to and from the final destinations. The driverless cars will work with and continually enhance their models of how animals, bicyclists, and human drivers (perhaps even specific drivers) behave in real-world situations so they will respond ever more appropriately.
Components. These seem to be essential, though how they are approached may vary:
- Car robotics. The machines that run the cars must be able to handle all the aspects of controlling speed, braking, changing direction, etc., rapidly enough to operate safely and effectively.
- Smart roads. Data on roads must be complete and up-to-date (beyond just a map, what Alexis C. Madrigal calls “ultra-precise digitizations of the physical world”), and markers for the cars (whether visual, electronic, or by wireless signal) must be embedded in the environment.
- Input/output devices. The car must sense the environment in a variety of ways and, possibly, be in communication with other vehicles and a central system.
- Logistics engine. The system will not be without constraints. Both the key factors (time, cost, capacity) and preferences (order of stops, type of vehicle, sharing, route) must be calculated and adjusted in the face of changes with the priorities of society and individual in mind. Cargoes spoil, meetings begin at set times, emergency vehicles need to reach people in jeopardy – a range of choices need to be made.
The impact.
- Loss of employment. Most people who drive for a living or to supplement their income will need to find other work.
- Energy conservation. Sharing, an end to traffic jams, using appropriately sized vehicles, and optimized routes are some of the ways energy will be conserved.
- Resource optimization. Some estimates say private cars (expensive investments in capital) are idle 90 percent of the time. Also, because people need to maintain stopping distances, the use of roads is well below potential capacity.
- Convenience. Having vehicles on demand, without maintenance concerns, will reduce the obligations of drivers.
- Enablement. Many people can’t drive or can’t afford cars. The system envisioned would put more jobs and resources within reach for those without cars. For an aging population, such a system could extend independent living.
Does this meet Stephenson’s challenge?
- Achievable. The intermediate may be achieved before 2024, depending upon social factors, including legislation. Things could go wrong if adoption is rushed and driverless cars are considered to be unsafe. Advances in technology, even to achieve the 2054 scenario, appear to be feasible. The full system implies an enormous capital investment. Who pays? Today in the US, over 65,000 bridges are in need of repair.
- Career objective. Is this big enough to take decades? I have real doubts about this. Where is the long-term inspiration for a graduating engineer?
- Practical purpose. Yes.
- A technological fix for a major human problem. From greenhouse gasses to access to services to resource conservation – the list of benefits is long.
I think, doing as the Atlantic article suggests, “making the world legible to robots” and “turning the physical world into a virtual world” will reach beyond transportation systems. Recent work in mapping the interiors of buildings suggests robots will be able to take the lessons of driverless cars into businesses, hospitals, schools, and homes. That and other extensions of this emerging transportation model may be what turns an inevitable project into a “big stuff” inspiration.
Is this a peek at the car of the future? http://techcrunch.com/2014/05/27/google-x-introduces-a-fully-self-driving-car-sans-steering-wheel-and-pedals/