PNS-MI stands for Peripheral-Nervous-System-Machine Interface.
A PNS-MI is a Human-Machine Interface which gathers signals from the subject; we concentrate upon non-invasive PNS-MIs, which means that we read muscular activity, skin deformation, ultrasound images, pressure footprints, accelerations, positions, torques, etc. These interfaces are often the only sensible way for the disabled (amputee, sufferer of neuropathic pain, stroke patient) to timely, naturally, dexterously control a robotic artifact – a prosthesis, an exoskeleton, a virtual reality environment, a computer.
These biological signals are coupled with target values (forces, positions, torques, kinematic configurations, etc.) obtained from other sensors or from synchronised stimuli, using machine learning techniques. The model so obtained can then in principle be used to predict the target values in the large.
The main PNS-MIs so far is surface electromyography, usually interpreted with thresholds or via classification; but one of our aims is that of designing, testing and implementing novel PNS-MI interfaces; we are also interested in new machine learning techniques, or new application of existing ones; last but not least, we design novel experimental methodologies, seek novel fields of application and bring together doctors and hospitals, engineers and mathematicians.