Electro Muscle Stimulation (EMS) provides rich opportunities for computer human control. Recently, EMS has been used in HCI to provide muscle-propelled force feedback and walking directions, to confer affordances and to share muscle control, for example. Due to individual physiological differences, these EMS systems require time-consuming, manual, per-participant calibration. Furthermore, existing approaches have relied on large-size, uniform-shape electrodes that make targeting specific muscles difficult and result in unwanted movement artefacts. As a result of this requirement for calibration, existing research has typically relied on limited, gross-movement stimulation gestures, such as rotations of the wrist or leg.
In this talk, I will present our ongoing work on auto-calibrating EMS through muscle signal repetition; reading muscle activation patterns with electromyography (EMG) and writing back with EMS. We use EMG to infer locations of key muscle activity and use this to drive complex, multi-electrode, multi-channel stimulation. I will present our control system for an electrode-dense wearable sleeve that enables pose-aware, complex and natural gestures, and discuss future research opportunities in this space.