A three-year US$600,000 will help researchers develop intelligent assistive robotics as part of a new research project, Mobile Robot Manipulators for Learning and Executing Instrumental Activities of Daily Living.
For people with upper extremity disabilities such as a stroke, multiple sclerosis or other conditions, assistive robotics can help restore their independence in performing everyday tasks like eating, grooming and grasping objects. However, these devices have a hard time learning “new tricks”.
“They need a lot of demos to learn new tasks, and people with disabilities aren’t able to provide these demos,” said professor Aman Behal, an assistive robotics expert at UCF.
“If they can, they might not be the best demos for the robot to learn from.”
Currently, mobile assistive robotics can be mounted to the side of a wheelchair or can trail behind the individual, ready to assist when needed. These devices allow users to navigate their homes or travel to places like an office or local park, offering greater flexibility than stationary fixed-base robotics.
While the dexterity of these robots allows them to assist with complex tasks like brushing hair or cutting a piece of steak, they can be difficult for the user to control or teach. Assistive robots perform tasks better over time by mimicking the user, but in these situations, the user may not be able to complete the task correctly – or at all.
Behal and his team of researchers, including Clinical Associate Professor of Physical Therapy Morris Beato, and Professor of Statistics and Data Science Edgard Maboudou, will create intelligent and mobile robotic assistants that offer easier control and a better robot-human interface.
The project’s first year will focus on research and development, while the second and third years will incorporate testing and user feedback. Initial testing will be conducted with students, followed by studies involving adults with upper body paralysis.
“We will recruit adults between the ages of 18 and 65, bring them (into our lab) and essentially get the robots to do several tasks,” Behal said. “At the end of it, we’ll see how easy it was to interact with the robot.”
The researchers will survey the participants to find out how many times they had to tell the robot to perform a task, how long it took for the robot to complete the action, how many mistakes were made and how satisfied they were with the robot’s assistance.
The end goal is to develop smart robots that can give the user a greater degree of independence, leading to increased self-esteem and enhanced quality of life.
