Javascript is required to view this video

10:05 a.m., March 9, 2011----Sunil K. Agrawal, professor in the Department of Mechanical Engineering at the University of Delaware, is co-author of a paper that explores how infants can be trained to avoid obstacles using mobile robots with force feedback joysticks.

THIS STORY

E-mail

Print

Twitter

Co-authors of the paper include Cole Galloway, associate professor in the Department of Physical Therapy, and graduate students Christina Ragonesi and Chen Xi. This work is funded by grants from the National Institutes of Health and the National Science Foundation.

Video at right shows Christina Ragonesi with a young toddler as he navigates the obstacle course.

Agrawal's paper, entitled “Training Toddlers Seated on Mobile Robots to Drive Indoors Amidst Obstacles,” will appear in a summer issue of IEEE's Transactions on Neural Systems and Rehabilitation Engineering journal. The paper is currently available for electronic review and download as part of the IEEE Xplore digital library.

“This is a landmark study where a group of infants were trained to drive mobile robots within an environment simulating clutter within the workspace,” says Agrawal. “The technology and algorithms were developed in my laboratory by doctoral student Xi Chen and tested at UD's Early Learning Center, with the results displaying overwhelming support that this “assist-as-needed approach” yields faster learning than a conventional joystick.”

The paper explains that mobility is a causal factor in development. In typically-developing infants, for example, the onset of crawling and walking has long been associated with developmental changes in perception, cognition and socialization. Infants born with significant mobility impairments often experience developmental delays due to lack of environmental interaction.

Previous studies by Agrawal and his research team have shown that, with several months of training, infants can learn to drive directly to a goal using a conventional joystick mounted on a mobility device. Higher-level skills needed to navigate around obstacles or turns, however, are not learned in the same time frame.

The algorithms developed by Agrawal's research team employ haptic rendering -- or sensory touch -- to speed learning in this young population by creating a force field on the joystick. If the child steers the joystick outside a force tunnel centered on the desired direction, the driver experiences a force bias on the hand. This feedback provides helpful sensory cues that quickly train young drivers.

During the study, ten typically-developing toddlers with an average age of 30 months were trained to drive a robot within an obstacle course. A toddler unable to walk independently due to spina bifida was also taught. Results on robot position and travel time were recorded, as well as the number of obstacle collisions.

The group study results indicated that the force field algorithm helped very young children learn to navigate and avoid obstacles faster, more accurately and with greater safety. The results from the child with spina bifida showed positive effects during practice with a child with mobility impairments. Study subjects were retested one week after training and early measures indicate the toddlers retained the learned behaviors at least one week following training.

“In the future, we hope to extend this novel application of technology and training to allow young children with special needs to explore and acquire other functional skills using power mobility devices in real environments such as their home or classroom,” concludes Agrawal.

Article by Karen B. Roberts