Based on the linked source; automatically prepared and checked against the original report.
Study objective
Researchers developed and evaluated a real-time electromyography (EMG) wheelchair-control framework for people with movement disabilities. The system used signals from either upper-arm or neck muscles, with the study investigating feasibility for assistive mobility applications rather than establishing a treatment or clinical recommendation.
Methods and evaluation
The researchers used feature optimisation to identify signals relevant to movement recognition, then trained and tested the model in offline and real-time settings. Feasibility measures included classification accuracy, repeatability, response latency and preliminary testing with two people with disabilities. The abstract describes the research record, not a full-paper review.
Reported performance and limitations
With a subject-independent approach, upper-arm control reached 95% real-time accuracy, while neck-muscle control reached 98.6% among able-bodied participants. Two disabled participants—a transhumeral amputee and a person with mobility impairment from poliomyelitis—achieved 100% neck-control accuracy. Average response latency was 430 milliseconds for upper-arm movements, 360 milliseconds for able-bodied neck control and 400 milliseconds for disabled participants. The authors reported feasibility in a controlled environment, while calling for larger, more diverse cohorts to assess generalisability and broader clinical applicability.
