Machine Learning: The Wake of Prosthetics

By StartUp City | Thursday, March 21, 2019

Most people who suffer the partial or total loss of the motor skills of the hand report a drastic decrease in the quality of life due to the consequent inability to perform many daily activities. Performing tasks that are often taken for granted, such as buttoning a shirt, using the phone, or grasping cooking or eating utensils becomes frustrating or nearly impossible due to the reduced grip strength and poor hand motor control that affects these people. A group of innovators from South Korea has developed a soft wearable robotic system with the purpose of helping these people to grasp and release objects in their vicinity.

This wearable robotics glove is in all aspects similar to a normal glove. It is composed entirely of fabrics that intelligently deform due to the action of compressed air applied to special and very thin chambers hidden in the glove, appropriately arranged in layers. The software applied here detects attempts to grasp an object through a camera feed by assessing the distance to the object and peculiar arm movements. Once the software has determined whether the user wants to grasp the object, soft actuators can be activated to provide the user's fingers with an adequate amount of assistive force. The system also includes a computer to enable the machine-learning algorithm to work and a module of actuation to help move the hand robot. The software will not be able to lend a hand to a person if the object is obscured from the digicam or its diverse viewpoint. The algorithm has to be improved by incorporating other sensor information or other existing methods of intention detection, such as using an electromyography sensor or eye gaze tracking.

This glove could be marketed in the future to improve the quality of life of motor handicapped people. Losing hand mobility can make daily tasks difficult or impossible, and the development of assistive technologies could substantially improve the quality of life. Currently, the technology has some limitations. The current system is a prototype, and the researchers want to miniaturize it to make it easy for a patient to carry. The technology is designed for use in patients with hand mobility impaired, such as patients with spinal cord injuries, stroke or cerebral palsy.