This article is more than 1 year old

DARPA: Send limbless troops back to war with robo-arms

Brainplug cyborgs to get superhuman reaction times

Radical Pentagon boffins have hit upon a new plan which will assist the USA in its efforts to meet the demand for combat troops in the expanding Wars On Stuff. Disabled servicemen who have lost arms or legs in the fighting thus far will be swiftly equipped with highly capable, ruggedised robotic replacement limbs and sent straight back into the fray.

In a government solicitation paper issued this week, we learn that:

Many wounded warriors with upper-limb amputation would like to return to their units in the field. However, in order to do so they must at least demonstrate the ability to field strip and reassemble their weapons ... [Bids from robotic limb designers must] adequately reflect potential clinical use initially by young and highly active wounded warriors (ie athletic individuals that strive to regain a high level of limb functionality).

The current state of the art in prosthetic limbs has been driven forward seriously in recent years due to the flood of young amputees arriving in Western-world hospitals as a result of the wars following 9/11. Some robo-limbs now have as many degrees of freedom, as much manipulative precision and potentially as much strength as the fleshy arms of fit, muscular soldiers.

This has called for some innovative portable power sources, including the use of a steam-powered arm running on potentially explosive peroxide fuels generally used only in torpedoes or rockets.

The problem now lies in control of these powerful limbs: researchers have struggled to create a viable link between the owner's motor-control centres and the machinery. This is what the new military Reliable Central-Nervous-System Interfaces (RCI) project is intended to address.

Such a brain-plug interface has long been sought: it would offer many benefits besides allowing troops to be swiftly returned to the fray after suffering disabling injuries. It could be used to restore function to undamaged limbs crippled by nervous-system damage, or serve as a hugely improved method of controlling computers or other machinery. Ultimately it might achieve one of the longest-held dreams of human science: the ability to place titanic, powerful robot juggernaut vehicles or colossi under the command of disembodied brains in bubbling jars.

So far, though, performance of prototype direct interfaces has been spotty and highly unreliable. Even where they work they tend to need frequent re-tuning and calibration, or involve tricky penetrations of the central nervous system's various biological protections which would present a terrible risk of infection or other problems under battlefield conditions.

Nonetheless we read:

Despite the above-mentioned challenges, recent technological advances in many areas (e.g., tissue-response-mitigating invasive cortical interfaces, high-density ECoG/epidural electrode arrays, non-contact EEG electrode arrays, multi-scale/multi-modal motorcontrol information acquisition, and robust motor-control intent-decoding algorithms) provide support to the following hypothesis: it is now possible to develop high-channelcount CNS interfaces that can reliably provide the amount of motor-control information needed for amputees to accurately and quickly control many-DOF prosthetic limbs for the remainder of their life at a performance level allowing them to perform activities of daily living both in military and civilian life.

The warboffinry chiefs of DARPA - for of course this programme could come from no other agency - say that they want a system that will last at least 70 years, offers at least 22 simultaneous motor-control outputs (about what one needs for managing a normal human Mark One arm and hand assembly) and has a Billy the Kid-like latency of less than 10 milliseconds*. They have $18m to spend in this, the latest quest for a proper brainplug interface system.

Sadly for any of our readers who may be lacking a limb or two - or have other reasons to fancy a 22-channel brainplug control socket - we have to offer the standard caveat on DARPA stories. The famous agency is specifically tasked with high-risk R&D - that is, R&D which is unlikely to pan out. The great majority of DARPA projects either fail or produce something unexpected: the classic example of the latter being the internet, produced by a DARPA programme intended to enhance military communications, which went out of control and mutated into the mightiest archive of pornography that human civilisation has ever known.

Thus the chances are that RCI won't turn out anything very useful. Even if it pays off, the odds of successful monkey-butler technology or something else unexpected are probably just as good as those for proper robo-limbs at last.

But at least someone's trying. You can read the RCI soliticitation in pdf here. ®

Bootnote

*For context normal human reaction time is around 200-250 milliseconds. It would seem that a robo-arm might make a soldier significantly quicker on the trigger than he was before being injured.

More about

TIP US OFF

Send us news


Other stories you might like