By now you know Google is doing research into making cars drive by themselves. This continues research done earlier at Stanford, Carnegie Mellon, and other places. If you really want to see innovation, watch the interview I did back in 2007 with Mike Montemerlo. He was the artificial intelligence guy behind Stanford’s DARPA challenge team (they won that contest back in 2005 and have been competitive ever since).
I remember thinking that Mike was doing the coolest stuff and that of all the geeks I met he had the greatest chance of really changing how the world works. After all, most of us spend hours per week in our cars, especially in California where public transportation options suck (I could take a bus to San Francisco, for instance, but I’d have to switch routes half a dozen times and it would take three hours, instead of the 45 minutes it takes driving.
So, why is Google doing this research?
Why is IBM, a few miles away, doing research into atomic structures?
A few weeks ago I was in Andy Wilson’s lab, talking with him about LightSpace. He asked me not to film, but you can see the video of his lab here. Who is Andy? Well, he was the researcher that came up with the ideas that turned into Microsoft Surface. He is doing some of the coolest research into futuristic computing technologies that I’ve seen anywhere. You really need to see Andy’s video (sorry, you’ll need Microsoft’s Silverlight loaded to view this video):
At IBM they let me move a single Iron atom across a piece of copper. And, at Stanford they are doing all sorts of research, including into weird flying robots that can be used for a bunch of different things (heck, one attacked me).
So, what are these folks telling me they do pure research for?
1. To figure out what’s needed from infrastructure in the future.
2. To discover how humans will use technology and how the technology will need to adapt to the humans.
3. To figure out how to affordably build futuristic tech.
4. To help researchers connect with each other and push the state of the art further.
5. To write up patents that will be the economic engines of their sponsoring companies.
6. To claim ground as “world leader” in a certain tech.
7. To build an ecosystem that turns into an industry (look at Boeing’s 787, for instance, and how its parts come from companies all over the world).
So, in that light, lets look at Google’s new self-driving car. Will it lead anywhere?
Here’s some things that research could lead to:
1. Better turn-by-turn directions for cars.
2. Better real-time mapping information. My Toyota already tells me whether there’s gas stations, hotels, or fast food at the next freeway exit. I imagine Google’s cars can tell you a lot more than that.
3. Better road design. Always look for unintended consequences of tech research. Remember, the Web came out of CERN who was smashing atoms. Google’s cameras can pick up confusing road design and map it for crews to fix, leading to safer roads for us all. My Toyota has trouble discerning old paint, for instance. Will these technologies be able to tell road managers when it’s time to repaint lanes?
4. New kinds of traffic controls. Listen to Mike and he’ll tell you the algorithms he had to write to get robotic cars safely through intersections. Will this new car research inform us as to the kinds of traffic controls (signs, lights, etc) that we’ll need in the future? Will it help us build new light timing systems to help traffic move smoothly?
5. New kinds of 3D control surfaces. These cars are building detailed 3D maps of the space that they travel through. Will those new surfaces need new controls? Of course! Second Life isn’t good enough for the data these cars are collecting.
6. New kinds of infrastructure. Each time these cars drive they gather gigabytes and, potentially, terabytes, of information. Now, imagine millions of cars that all could report home about what they were seeing on the road. That will require building new datacenters, new ways of filtering that information, and new database technology that can handle huge amounts of real-time info.
7. New kinds of human/computer interactions. If your car is going to drive itself, at least some of the time, you’ll need new kinds of input methods (voice? Touch? Gestures?) and they will need to be tested out on a lot more than 140,000 miles that these Google cars have driven so far.
Anyway, there’s lots to learn. I’m glad that the tech industry is taking on research in these, and other, areas. Even ones where people might say “that’s lame.” I remember when my friends told me Twitter is lame, but now it’s very important way to get the news.
If you are doing world-changing technology research I would love to come and visit you. Send me email at firstname.lastname@example.org or call me at +1-425-205-1921
Why else do research? Comment here and let me know if you are doing technology research and why you’re doing it.