HAL 5, design issues, continued

I have often wondered how much potential myoelectric relays really have. In the case of the HAL suit the idea is this: The sensors on the skin detect the electrical charge in your skin caused by the nerve sending a signal to your leg muscle to move your leg. The HAL suit then move its leg, and your leg moves with it.

That's pretty amazing.

But the human skin is a complicated organ. Let me give you an example. If you were (go with me here) to plug in an oscilloscope, and plug a wire into the input jack, then hold the other end of the wire, the oscilloscope would be flat. But if you then turned on the flourescent lights in the ceiling, the oscilloscope would immediately detect a 60 Hz frequency signal, somehow coming from your skin. The reason for this is that your skin acts as a signal receiver for the 60 Hz signal being emitted by the electrical current that powers the lights, the same way a radio antenna detects the FM or AM radio station signal traveling through the air.

So how does the HAL suit discern what is a nerve signal and what is background noise? If your skin is getting constantly bombarded with thousands of signals at different radio frequencies, all of which it picks up and amplifies, how does the HAL suit cut that out? And what about static charge? If you scrape your feet across the floor, does it build a static charge which is absorbed by the sensors, causing a false positive? What happens when you get sweaty, and your body becomes covered in a glistening layer of conductive liquid, spreading and dissipating signals from your nerves? And what about when you just flex your muscle, but don't move your leg? I'm sitting here and I can flex my thigh muscles. Would the HAL suit try to move my leg, or does it somehow know better?

Over at the RIC, a similar method is being used to control artificial limbs for amputees. Electrical sensors are placed on the pectoral muscle of the amputee, who then learns, slowly but surely, to control a simple mechanical arm using careful muscle activation in his/her pectoral.
However, this signal is highly degraded. Direct nerve-to-wire connections would be more advantageous, could they be devised.


_