Self-driving cargo—I can see it all now
When I think about self-driving cars, my imagination naturally jumps ahead to self-driving cargo. This is even easier when I see pedi-cargo carriers, as I did Wednesday morning (March 9) on West Broadway Avenue. (It’s so easy to imagine them with no one pedaling!)
Three distinct areas of research and development: self-driving cars, delivery drones and electric vehicles will (I believe) come together in the form of electric-powered, self-driving, delivery vehicles.
With compact computer guidance systems and electric propulsion taking the place of both drivers and internal combustion engines, such cargo robots could be little more than boxes on wheels (the little more being cameras, sensors and, perhaps, soft rounded corners).
The drive to automate whatever can be automated
Google’s cutesy self-driving cars may be the face of autonomous vehicle technology but as much money as there may be in selling cars that can drive their owners to the corner store, there’s probably far more money to be made replacing the fleets-worth of delivery drivers and taxi drivers and and bus drivers and garbage truck drivers and dump truck drivers and so on.
Computer automation began in the 1950s by tackling the most mindlessly repetitive and predictable of tasks. The simplicity of what could be automated, necessarily mirroring the state of computer science.
In the same way and for the same reason that automation finally eliminated the need for human elevator operators just before I was born, autonomous vehicle technology aims to do away with all the driving jobs that exist for no other reason than motor vehicles have always needed to have a driver.
In self-driving vehicles, computer science is betting that it can now tackle the repetitive complexity of logistics–of safely moving people and things from point A to point B. This is the sort of task where predictability increases and decreases inversely proportional to scale. The map is known. The route is known. What remains to be seen are all the little problems that will crop up at any given moment along the route.
The devil, as they say, is totally in the details.
However, as unlimited as the potential problems may seem, they really aren’t and 99.9 percent of the time they can be solved by drawing on an equally large and equally finite set of know solutions.
That’s the theory and fortunately for computer science human drivers have set the bar for acceptable rates of problem solving in traffic comfortably low.
Delivering the future in a familiar package
There are literally dozens, if not hundreds of projects working to develop workable self-driving delivery systems. Here are just a few.
On February 9, Google was awarded U.S. patent number 9,256,852 B1, which describes an “autonomous delivery platform”, aka a self-driving delivery truck.
In the patent, Google envisions a truck that would carry packages in a cargo box made up of individually secured lockers. Customers would be able to open the locker containing their package using either a PIN code or swipe card supplied to them as part of the shipping transaction.
In April 2015, the U.K. arm of Domino’s Pizza unveiled the fairly radical-looking “Domi-No-Driver” autonmous delivery vehicle–a motorcycle with a pizza compartment and round head and two “eyes”. In 2013, the U.K. Domino’s trialed delivering pizza using the DomiCopter helicopter drone.
Meanwhile, in October 2015, the parent Domino’s Pizza in the U.S showed off a new custom pizza delivery car—a modified Chevrolet Spark. Not actually self-driving but developed in conjunction with Roush Enterprises, the company building Google’s self-driving cars.
In November of 2015 Starship Technologies received a ton of press for its storage tote-sized self-driving delivery robot.
The Starship’s robot is designed to travel sidewalks at a top speed of 6.4 kilometres per hour. It is described as 99 percent self-driving, with a human operator standing by to pitch in remotely if needed. This is apparently similar to an approach patented by Google for getting its self-driving cars out of difficult spots. (The robots still need us after all!)
Hopefully self-driving cars never see Tron at a drive-in
As a total aside, there’s a bit of irony in the way that self-driving cars will finally turn a long-standing trope about computers inside out.
One of the first (if not the first) literary references to city streets resembling paths on a circuit board occurs in Thomas Pynchon’s 1965 novel The Crying of Lot 49.
The idea of computer architecture mimicking the real world culminated in the 1990s Vancouver-based CGI cartoon series Reboot, which was set in a richly-imagined digital/urban environment located inside a mainframe computer. But of course, Reboot owed its entire inspiration to Disney’s 1982 film Tron.
Remember how in Tron, and again in the 2010 sequel Tron Legacy, when the unseen computer users played computer games such as Light-Cycle? It was the actions of the users in the real world that drove the life and death game simulations inside the computer.
Well, self-driving cars are exactly the other way around. What the software processes simulate inside their digital world will actually drive users in the real world.
Thar’s funny but maybe it’s serious as well. Fifty years ago, there was no doubt that what went on inside computers wasn’t real and any comparison between the architecture of computers and the real world was meant as literary allusion and nothing else.
Today a great deal of important stuff only exists in the real world as a copy of something created inside a computer: movies, art, all sorts of written documents—sculptures even.
The idea that the software simulators of a decade ago may have evolved enough to drive the cars and potentially fly the planes all by themselves isn’t shocking—our relationship with computers has changed that much.
Maybe the once clean and certain line between the digital and the real is beginning to get a little fuzzy and maybe it will get a whole lot fuzzier.
Perhaps a hundred years from now things will have gotten to the point where people can seriously ask: Which came first, Earth, or Google Earth? Click the images to enlarge them.