Millions of people around the world have lost their upper limbs mainly due to accidents and wars. Recently in the Middle East, the demand for prosthetic limbs has increased dramatically due to ongoing wars in the region. Commercially available prosthetic limbs are expensive while the most economical method available for controlling prosthetic limbs is the Electromyography (EMG). Researchers on EMG-controlled prosthetic limbs are facing several challenges, which include efficiency problems in terms of functionality especially in prosthetic hands. A major issue that needs to be solved is the fact that currently available low-cost EMG-controlled prosthetic hands cannot enable the user to grasp various types of objects in various shapes, and cannot provide the efficient use of the object by deciding the necessary hand gesture. In this paper, a computer vision-based mechanism is proposed with the purpose of detecting and recognizing objects and applying optimal hand gesture through visual feedback. The objects are classified into groups and the optimal hand gesture to grasp and use the targeted object that is most efficient for the user is implemented. A simulation model of the human hand kinematics is developed for simulation tests to reveal the efficacy of the proposed method. 80 different types of objects are detected, recognized, and classified for simulation tests, which can be realized by using two electrodes supplying the input to perform the action. Simulation results reveal the performance of proposed EMG-controlled prosthetic hand in maintaining optimal hand gestures in computer environment. Results are promising to help disabled people handle and use objects more efficiently without higher costs.
© 2019 by Walter de Gruyter Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 Public License.