The DARPA Reliable Peripheral Interfaces program seeks to develop reliable in-vivo peripheral motor-signal recording and sensory-signal stimulating interfaces for robust, dexterous, and high degree-of-freedom prosthetic limbs. The goal of this JPL task is to demonstrate clinically viable algorithms for reliably decoding motor-control signals from detected peripheral signals, beginning with surface electromyography (EMG) signals from an array of sensors in an arm sleeve. JPL will develop real-time decoding algorithms to process spatio-temporal surface EMG patterns, including separation of muscle channels, tracking of individual muscle activity, and self-calibration algorithms for alignment on donning the EMG sleeve. The overall goal of this effort is to advance the field of signal decoding from peripheral signals to improve human-to-machine interfaces.
Christopher Assad - Jet Propulsion Laboratory