Video: Crowdsourcing App Helps Space Agency Improve Robots
A free iPhone video game app turns your Parrot AR.Drone into a simulated spacecraft, which you can use to simulate docking on the International Space Station. You get points for accuracy and speed, and the European Space Agency gets tons of data to help make better space robots. (Source: European Space Agency/Anneke Le Floc'h)
Excellent story and unique application for using mobile devices. I don't think I have the dexterity and coordination to make this work but I'm sure there is a whole generation of gamers than can help them gather data. Thanks.
Glad you enjoyed the story, Al. I thought it was a fun app, and also good to know that at least some space agencies are open to the crowdsourcing concept. I've heard of SETI's requests for help from millions of people with PCs, but not anything about NASA using the crowdsourcing approach. Anyone know?
Every time I see one of these stories, I can't help but think of an old Chicago-ism: "They're muscling in our rackets." Increasingly, I'm seeing a lot of tasks that robots can do more effectively than humans. And the kicker to this story is they now want all of us to help them learn. I know it's logical; it's all in the best interests of science and technology; it's probably helpful to mankind in a hundred different ways that I can't even imagine. Still, I have this niggling fear, and I know it's not the most enlightened view -- but they're muscling in our rackets.
Chuck, your comments make me think of black and white 40s gangster movies, many of them located in Chicago. Anyway, in this case I think teaching robots how to dock on the ISS makes more sense than trying to teach humans to do it. Of course, astronauts and people who want to be astronauts might not agree.
Ann, maybe there's a movie here -- a cross between 2001: A Space Odyssey and Public Enemy. As to it making sense for applications that humans don't want to do, I concur. From a completely logical standpoint, I know that robotics make sense in countless applications. Admittedly, my fear is illogical. Still, every time I read one of these stories...
Chuck, I understand the fear. I've read a ton of science fiction since age 11, so I probably have some of the same worries you do about the Robot Takeover and the War with the Machines (oops, wrong universe). At least I'll probably be long gone if that happens.
Relying on "crowdsourcing" is not a fast moving option. Depending if people are paid or not, work takes a while to push through. People either have to absolutely love the product or company or get paid to help a product along...
Or in the immortal word of one Internet meme "Ain't nobody got time for that!"
Cabe, if you're responding to my comment, "that didn't take much motivation," I meant that people who have already invested $300 in the AR.Drone--not small change if you ask me--therefore don't need much motivation to use it in this crowdsourcing app. You had said that getting people to participate might be problematic, especially if they weren't paid. But dedicated gamers like these don't need an excuse to play.
Al, one aspect of AI that these researchers are trying to fulfill with crowdsourcing is to gather lots of data (lots of people in the crowd). Even if you were not very skilled at it, that data might be useful as well.
The main point of crowdsourcing is big numbers, so lots of data. In this case, ESA has designed the app to motivate players to do an excellent job, thus providing the type of data they want, i.e., what are the best ways to approach Object A reliably with millions of copies of Object B, in turn learning methods that can be taught to other robots.
In addition to to using crowd sourcing to improve their design, ESA also builds public awareness to their space programs (which could ultimately lead to more favorable funding for certain programs by the public and their government).
It seems that somehow the data created as various folks attempt to dock at a simulated space station is somehow going to be useful. Right??? But the game will certainly obtain lots of data, for sure. But the main value of data is created when it is condensed into knowledge, and the value of knowledge is that it can lead to insight and understanding. Exactly how that happens in this situation is not completely clear to me just yet.
William, there's more detail about what the researchers are doing, and plan to do, in the links we gave in the article, including a list of references at the end of the AI project article There's also a list of publications by the Advanced Concepts Team here: http://www.esa.int/gsp/ACT/publications/index.htm
Ann, thanks for pointing out that thye links had additional useful information. Unfortunately for me, I guess, is that I seldom follow links placed in articles, partly because some of those links have been quite slow in the past.