Human prosthetics in the near future will require communication with the nervous system. It is therefore necessary to create implantable devices that can read electrical information sent by the brain. The Analog Design Team at the University of Texas at Dallas has developed a prototype chip that can be powered wirelessly and communicate with an external interface through the skin, eliminating the need for battery maintenance and skin breaks for communication. It is designed to be connected to the peripheral nerves that previously controlled the lost limb, thus minimizing rehabilitation. This thesis describes an adaptive neural spike threshold sensor used on the chip to determine an appropriate threshold level for nerve spikes. The RMS level of the noise in the nerve signal is used to set the threshold. The sensor dissipates 2 muW, has an area of 1.77mm2 and incorporates a novel charge pump based approach to setting the threshold.
展开▼