Researchers at Florida Atlantic University have developed a soft robotic glove that can assist stroke patients to relearn how to perform dexterous tasks with their hands. The glove contains soft actuators that helps patients to move their fingers as they perform tasks, and sensors that help to create tactile sensations. So far, the researchers have focused on using the gloves to assist patients to play music. The team used machine learning to train the gloves on a simple tune, and then when users play the tune the gloves can provide feedback on where they went wrong. This proof-of-concept shows that the technology could assist patients in relearning a wide variety of tasks that require significant dexterity.
After a stroke, patients can experience significant impairment and often require physical rehabilitation to allow them to relearn tasks, from speaking to walking. Tasks that require significant dexterity, coordination, and very fine movements, such as playing the piano, can be particularly difficult to relearn. To address this, these researchers have created a soft robotic glove and accompanying AI system that is specifically targeted at enabling stroke patients to relearn such dexterous activities.
The glove contains a series of soft actuators that assist with bending and extending fingers. However, these actuators are designed to make the patient’s intended movements easier, rather than control their movements outright. The glove also contains 16 flexible sensors in each fingertip, that enhance the tactile sensation that a wearer experiences when touching a surface.
“While wearing the glove, human users have control over the movement of each finger to a significant extent,” said Erik Engeberg, a researcher involved in the study. “The glove is designed to assist and enhance their natural hand movements, allowing them to control the flexion and extension of their fingers. The glove supplies hand guidance, providing support and amplifying dexterity.”
The system uses AI to inform a training component. The researchers used machine learning to teach the system the difference between correct and incorrect playing of a simple tune. Then, when a user plays the tune, the system provides feedback on where they may have gone wrong, assisting them in improving.
“Adapting the present design to other rehabilitation tasks beyond playing music, for example object manipulation, would require customization to individual needs,” said Maohua Lin, another researcher involved in the study. “This can be facilitated through 3D scanning technology or CT scans to ensure a personalized fit and functionality for each user. But several challenges in this field need to be overcome. These include improving the accuracy and reliability of tactile sensing, enhancing the adaptability and dexterity of the exoskeleton design, and refining the machine learning algorithms to better interpret and respond to user input.”
Here’s a Florida Atlantic University video about the technology:
Study in journal Frontiers in Robotics and AI: Feeling the beat: a smart hand exoskeleton for learning to play musical instruments