Funding Organization:
National Science Foundation
Funding Amount:
$480,000
Funding Period:
2019-2023

Over 100,000 people in the United States have upper limb amputations. Such amputations are usually associated with substantial disabilities. Activities of daily living may no longer be possible or require additional effort and time. To restore functional ability as fully as possible, amputee patients use upper-limb prostheses controlled by Electromyography (EMG)-based human-machine interfaces (HMIs). However, over 50% of prosthetic users report device dissatisfaction and abandonment due to frustration in use. In order to increase accessibility and utility of EMG-based assistive HMIs, it is critical to investigate the cognitive workload associated with using these systems for supporting motor skill rehabilitation and activities of daily living. In addition, with differences in individual muscle make-up and control, it is necessary to customize EMG-based assistive HMIs to particular patient conditions. However, at this time, there exists little to no general guidance on what interface control features may be more or less conducive to supporting learning and performance of psychomotor tasks. By testing various control interface prototypes in a validated motor skill learning simulation as well as high-demand, real-time tasks, specifically simulated driving, the investigators seek to provide design guidance for engineers to develop EMG-based assistive HMI technologies for various applications. This project also provides education and training for young industrial, computer science, and biomedical engineering students in assistive technology design and development.

The technical aims of the project are divided into three thrusts including (1) identifying cognitive load costs of EMG-based HMIs using computational cognitive performance models and machine learning algorithms – The developed cognitive workload models will be used for predicting learning potential and performance outcomes of variations of control interface designs, (2) demonstrating fundamental motor skill training through integration of the EMG-based HMI with virtual reality (VR) simulations – In particular, the investigators are interested in assessing whether the difference in cognitive load of using EMG-based HMIs translates to variations in learning potential and retention of psychomotor task skills, (3) translating fundamental motion components trained in the VR psychomotor test to a real-world application by demonstrating the possibility of operational driving control with an EMG-based HMI in a high-fidelity driving simulator – This thrust provides empirical validation for the findings regarding differences in cognitive load of using EMG-based HMIs, as well as the potential for increased learning and operational task performance.