OpenSense: An open-source toolbox for inertial-measurement-unit-based measurement of lower extremity kinematics over long durations.

Only 11 healthy young adults were studied, mostly male and with a mean body mass index in the normal range, so it does not show how the workflow performs in older adults, people with stroke, or patients with movement disorders. All testing was in a laboratory with electronic equipment and ferromagnetic materials, including in-ground force plates, so magnetic disturbance may be different in clinics or homes. The validation lasted 10 minutes, not hours or days, so longer-term drift remains uncertain. Seven of 11 subjects had at least one sensor excluded by the pre-screening, and one subject was omitted from a walking-plot analysis, so not all data were usable. The researchers used optical motion capture data to register the model to the IMUs, which is not realistic for natural environments. Kinematics were computed offline, not in real time, and the workflow was tested on healthy adults performing prescribed lab tasks rather than clinical outcomes. Hip rotation was the least accurate joint angle, with a median RMS difference of 12°, and foot sensors were most affected by magnetic disturbances. The instructed neutral calibration pose differed from the subject's actual pose by up to 20.1°, especially hip rotation.

The metadata lists NIH extramural and non-U.S. government research support. The supplied article text does not include a separate author conflict-of-interest declaration.

Do not read this as evidence that wearable sensors are ready for patient gait assessment. It tested 11 healthy adults in a lab, excluded some sensors due to magnetic disturbance, and computed kinematics offline.