OpenMyoControl system enables muscle-reading control for tetraplegia with any sensor placement
Ultrasound-based control works anywhere on body with <5.5% tracking error after 3 pose calibrations.
A new breakthrough in assistive technology allows people with tetraplegia to control devices using ultrasound signals from any muscle—regardless of where the sensor is placed on the body. The system, developed by Gavin Sueltz and colleagues as part of the open-source OpenMyoControl project, uses sparse optical flow tracking on sonomyography (SMG) data to decode muscle deformations into continuous control signals. Unlike prior SMG systems that required extensive per-user, per-location training, this approach needs only minimal calibration: 3 pose definitions for 1-DOF control, and a short computer-aided calibration for 2-DOF control.
In a study of 3 cervical spinal cord injury survivors and 6 uninjured individuals, the system was tested across 6 sensor placements spanning the arm, neck, and upper torso—even including passive tissue motions. All participants achieved continuous 1-DOF cursor control with less than 5.5% tracking error on at least one placement, with many achieving under 4% across multiple locations. For 2-DOF, participants could modulate cursor position, and several completed a drawing task—the first demonstration of location-agnostic multi-DOF continuous SMG control. The team has open-sourced the algorithms on SimTK to accelerate development.
- System works with any sensor placement on arm, neck, or torso — no per-location retraining needed.
- Achieves 1-DOF control with just 3 calibration poses; 2-DOF control demonstrated on tetraplegic users.
- All 9 participants (3 with tetraplegia, 6 uninjured) achieved <5.5% tracking error on at least one placement.
Why It Matters
Rapidly calibratable, placement-agnostic muscle control opens practical assistive tech for spinal cord injury survivors.