In the wake of the success of Predator and Packbot in Afghanistan, there's a rush to create variants, particularly designs useful in urban warfare. What we see in Iraq - probably in a few weeks' time - won't significantly differ from Afghanistan. There's just not enough time to create new platforms, or build and deploy lots of the current ones. There may be a few variations on the existing themes. Predator has already been armed with Hellfire missiles, and scored the first UAV kill in Yemen. There are publicly disclosed proposals to directly connect the Predator output to combat helicopters for use as a battlefield scout, and to load on an electronic intelligence (ELINT) pod for jamming or targeting of air defenses.

The biggest transformations lie in the future, but not far off. Winds of Change quotes and points to a JED Online article (registration required) describing the Loitering Electronic Warfare Killer (LEWK) concept: a swarm of small sensor/killer aerial bots designed to assist air space penetration. Rather than control a single UAV platform, the attacking helicopter or fighter would unload a pack of 18 to 50 ($100,000 each) LEWKs to precede it into harm's way. Obviously, these aren't going to be individually controlled. They are going to be semi-autonomously pursuing goals set by the operator. That's the next big change: robots that perform as teams. Don't think of replicants, think about fighting, flying ants with specialized forms and roles, coordinated in real time.

The cited project is just the tip of the iceberg, both in the military and other domains. Here's a PDF of a study on genetic algorithms for evolving surveillance behavior in micro-aerial vehicles. And a lab pursuing sensor fusion for urban search and rescue, but read to the end of the blurb:

...we concentrate on the coupling of perception (especially from multiple sensors) with action through intelligent sensor fusion. Our primary test domain is mobile robots for Urban Search and Rescue (think Oklahoma City, Kobe, other disaster sites... ). We are interested in the use of cascaded, heterogeneous robots teams (aka marsupials) to cooperatively accomplish complex missions, especially using distributed and surrogate sensing. Funding sources for the lab include DARPA, ONR, NSF, and companies such as SAIC and K-Team.

And then there are the team oriented robot competitions, with entrants such as CMU's premier team, which uses a set of cooperating AIBO variants to play robo-soccer. (I saw this one demonstrated at DARPATech last year. Must admit it was a hoot watching a bunch of colonels and better in full military regalia keeping a stiff upper lip while a bunch of dog-bots kicked around a little orange rubber ball.)

Add it all up. Whether civilian or military, even a nominally humanoid 'bot is likely to be a specialized worker or warrior member of a robotic team.
5:11:42 PM    

Just like 802.11b has a industry association called WiFi to drum up support and do things like certify compatibility, now there's a baby WiMax Forum that wants to do the same thing for 802.16 broadband wireless access. Good luck to them, but a couple of caveats at the same time. While the WiFi gang has dozens of heavy hitters as members, WiMax can now muster less than one dozen. Of them, only three could be considered heavy: Nokia, Agilient and Hughes. All of them suppliers of high priced gear to incumbent carriers. When there's a longer list, including some upstarts with no vested interests in incumbents or high prices, like (for instance) Acer and LinkSys over at WiFi, then we'll know there might be some momentum about to build, particularly for the 802.16a unlicensed flavor of wireless broadband.
2:04:18 PM    

NTT DoCoMo will pick up 50% of the tab for product development by Fujitsu, Matsushita, and NEC, says a European media trade rag. Now, it's not at all unusual for carriers to subsidize the end user costs of handsets, but bearing the capital costs of development is not common, particularly when aiming at a growing (3G) market. The article mentions an ill-defined sharing of copyrights and patents in the products (which are supposed to be standards based). That looks like a fig leaf. One of two things is likely going on here: Slipping some cash to local champion companies to bolster their export capabilities against everyone from Nokia to Ericsson and Samsung, or a real need by DoCoMo to get back vendors' attention from the CDMA-based photomail successes of J-Phone. If the former, it's business as usual. If the latter, DoCoMo's position and/or vendors' finances have eroded more than externally visible: Usually DoCoMo just has to say 'jump' and the vendors ask how high.
11:28:41 AM