Prosthetics enable mobility and independence but fitting the apparatus and helping recipients learn to use an artificial arm or leg can be obstacles in the way of success. Powered prosthetic legs, for example, typically require “tuning” by trained clinical practitioners before patients can walk with ease. We’ve written about adding touch sense to prosthetics and training patients to control their prosthetics with their minds.
Researchers at three universities have developed artificial intelligence-based lower limb prosthetic tuning that helps patients walk in minutes with no discomfort. The team from North Carolina State University, the University of North Carolina, and Arizona State University published a study describing their work in IEEE Transactions on Cybernetics. According to Helen Huang, a co-author of the study and a professor in NC State’s and UNC’s Joint Department of Biomedical Engineering, a human clinician modifies just a “handful” of parameters so a patient can use a powered prosthetic knee walking on a level surface. This tuning process can take hours. In contrast, the intelligent prosthetic modifies 12 parameters and can have patients walking in about 10 minutes.
Further work remains to be done on the project, including walking on inclines and declines, as well as up and down stairs. The group also hopes to speed the tuning process even further by focusing on parameter combinations that have a greater likelihood of success. The researchers also want to design a system that can adapt to the patient’s gait and preferences as well as learn over time for even better comfort and efficiency.
