Reinforcement Learning ‘Tunes’ Robotics Prosthetics in Minutes
Researchers from Arizona State University, the University of North Carolina and North Carolina State University have discovered a new intelligent system that depends on reinforcement learning to tune powered robotic prosthetic knees.
This new development will allow patients to start walking comfortably using their prosthetic equipment in minutes, as opposed to the hours it would take if the gadget was to be manually tuned by a trained clinical practitioner. This invention is the 1st system to rely only on reinforcement learning to tune robotic prosthesis.
According to a statement by North Carolina State, the new tuning device tweaks 12 distinct control parameters to accommodate the specified patient while addressing prosthesis dynamics like joint stiffness through the whole gait cycle.
Usually, a human medical practitioner would work with the prosthesis’ user to adapt several parameters. This process takes hours to complete. The new invention, on the other hand, relies on a computer program that avails a type of machine learning called reinforcement learning to modify all 12 parameters instantaneously.
This permits the patient to use the powered prosthetic knee to walk on a level surface in less than 10 minutes.
“We begin by giving a patient a powered prosthetic knee with a randomly selected set of parameters,” said Helen Huang, a co-author of a paper on the work and a professor in the Joint Department Biomedical Engineering at UNC and NC State. “We then have the patient begin walking under controlled circumstances.”
“Data on the patient’s gait and the device are collected via a suite of sensors in the device,” she added. “A computer model adapts parameters on the device and compares the patient’s gait to the profile of a normal walking gait in real time.”
Even though all the work is currently executed in a controlled, medical setting, the goal of the researchers is to invent a wireless version of the system. This will permit the device to continue fine-tuning itself as the user tackles his or her daily activities and surrounding conditions.