Sun

Jan 21
2007

Tim O'Reilly

Tim O'Reilly

DARPA LoCo - Location Anticipation

Ross Stapleton-Gray wrote:

DARPA has just come out with a new solicitation, "LOCO" (for Location and Connection Aware Content Pushing.)

In a nutshell, it posits that, while caching, P2P collaboration in data hosting, etc., can efficiently get content across the Net, what will be needed by the U.S. military is a network that anticipates what's needed in advance of requests. Some of that's intuitive (e.g., detailed tactical maps of areas being entered, order of battle for units likely to be encountered, etc.), but I think their targets are wildly optimistic, e.g. (by the end of less than three years of development) demonstration of a system scalable to more than 100,000 machines, able to deliver at least 50% of information required in advance of a request at least 80% of the time, while delivering unused information less than 10% of the time.

DARPA plans to make two to four awards to companies/teams, across two tasks: architecting such a system, and test & evaluation, including construction of a testbed.

I'm working with a company that is interested in the test & evaluation portion of the solicitation, and seeking teaming partners. There will also be a Proposer's Day Conference on Feb. 27, in Arlington, VA.

This idea of anticipation as the next step in location awareness is a very interesting one, especially considering Jeff Hawkins' idea of the brain as a prediction engine.


tags: geo  | comments: 5   | Sphere It
submit:

 
Previous  |  Next

0 TrackBacks

TrackBack URL for this entry: http://blogs.oreilly.com/cgi-bin/mt/mt-t.cgi/5181

Comments: 5

  Peter Cranstone [01.21.07 08:10 AM]

We currently have a solution which could provide the transport mechanism. It's a client server solution. On the server side is Mod_LBS (Location Based Services) which interacts with an ultra thin client on the target device (desktop, laptop, UMPC, cell phone) and allows you to send and receive GPS and Device information in real time over HTTP. It's bearer agnostic and will work with any communication medium that supports HTTP. We use encryption on the the target device to secure the content and another encryption method to secure it during transport. Depending on feedback from the GPS and or sensors in the target device it provides a real time feedback information system to generate real time content push.

More information on our web site (http://www.5o9inc.com)

Cheers,


Peter

  serge [01.21.07 12:20 PM]

Tim, I greatly enjoy your posts and I visit weekly. I have another comment: The human brain can predict with some degree of success because it is intelligent, not because it is a prediction engine. To say that intelligence is a prediction engine and that therefore this explains intelligence is a tautology, a circular reasoning, typical of Plato's cavern, as you had known some years ago (e.g., your PhD thesis). My favorite description of the essence of human intelligence is that it can somehow come up with new and more valid sets of hypotheses (about the observed reality). We still do not know how we do this, how we come up with better sets of hypotheses. When we know this then we may be able to make machines that emulate this kind of intelligence, and these machines will be intelligent, at least in the human sense, and maybe not in the dolphin or whale sense of the concept of intelligence.

  jimi [01.21.07 06:32 PM]

There is no Soviet Union anymore, but everybody remember those great victories and defeats. We trusted in idea and we made our history through great losses...
www.backinussr.com

  Ross Stapleton-Gray [01.21.07 10:42 PM]

As I mention, I'd be interested in talking with anyone interested in submitting a proposal, especially on the test & eval side of things, for either of the two solicited tasks. I'd have to say, though, the targets they're looking to hit seem wildly optimistic, particularly on the desire for a low "false positive" (information fetched but not needed). Think of any scenario where a military unit encounters an unexpected circumstance, e.g., a blown bridge; not only are you suddenly in need of a lot of new information on alternative paths, you're discarding (or no longer needing) a lot of information it was quite reasonable to expect you'd want, when a planned course of action is abandoned.


But there's another interesting wrinkle here: if the network gets that much more intelligent, it increases the risk of operational security liabilities... compromise the network, and you've access to highly sensitive plans and intentions. Or tweak a few bits, and watch the military's global neural synapses slow to a molasses pace...

  Robert D. [01.23.07 12:22 PM]

"but I think their targets are wildly optimistic, e.g. (by the end of less than three years of development) demonstration of a system scalable to more than 100,000 machines"

Never underestimate the power of a runaway military budget.

Post A Comment:

 (please be patient, comments may take awhile to post)






Type the characters you see in the picture above.