Pull the sensor (sight and hearing) processing out of the robot and comunicate with the sensors, via a wireless LAN (Like Honda's Asimi) with an IR backup. Then do the image and sound processing with a couple of Beowulf arrays, maybe even some limited AI on another. It's do-able with todays technology. Just needs better power options and lots of money.
Slashdot Mirror
Pull the sensor (sight and hearing) processing out of the robot and comunicate with the sensors, via a wireless LAN (Like Honda's Asimi) with an IR backup. Then do the image and sound processing with a couple of Beowulf arrays, maybe even some limited AI on another. It's do-able with todays technology. Just needs better power options and lots of money.