Inspired by online music streaming services, a team of engineers from the University of Michigan, Google, and Georgia Tech University have designed a very simple way to program their own settings when playing music using an exoskeleton to help them.
Of course, specialists point out that what is simple for the average user is more complex than it seems, while machine learning algorithms frequently offer assistance profiles that could be more comfortable for the person using the exoskeleton.
The user then selects one of the two profiles provided on each occasion, and the prediction system suggests a new help profile, which it deems would be better. This approach allows users to select the assistance provided by the exoskeleton according to their preferences, all using a very simple interface, which can be implemented on a connected watch or smartphone.
“It sounds like the online music service Pandora,” says Elliot Ross, one of the authors of the study published in the journal Science. Scientific robots. “You give it a review, either excellent or excellent, and it suggests a radio station based on your feedback. The idea here is similar, but it has to do with the ancillary parameters of the exoskeleton. In both cases, we create a model of the user’s preferences and use it to improve the individual’s experience.”
The team tested this approach with 14 participants, each of whom used a pair of exoskeletons strapped to the ankle, while walking at a fast pace of about 4 km/h. Volunteers could spend as much time as they wanted between two options, although they were limited to 50 options. Most participants chose the same helping profile, repeatedly, when they were on their 45th decision, we note.
After 50 rounds of decision-making, the study team began surveying users to determine whether the last help profile was truly the best, and evaluated it against 10 other random, plausible profiles. On average, participants chose the parameters selected by the algorithm about 9 times out of 10, highlighting the accuracy of the proposed approach.
“Using smart algorithms and a touch of artificial intelligence, our systems determine what users want with simple yes or no questions,” said Ong Hill Lee, lead author of the study. “I am excited that this approach will ensure that wearable robots will be comfortable and easy to use, bringing us closer to the day when their use will be an integral part of our daily lives.” »
The control algorithm helps manage four parameters of the exoskeleton: the amount of force delivered to assist the user, the delay between power peaks, and how the exoskeleton increases and decreases its assistance on either side of these power peaks. This approach is based on how the leg muscle adds strength to move us forward with each step.
According to Mr. Ross, only a few exoskeleton users can set their own parameters.
“In most cases, the control mechanisms are based on biomechanical or physiological results. Researchers adjust settings on their computers, which reduces users’ metabolic rate. Currently, this is the standard for evaluating and monitoring exoskeletons,” the specialist said.
“I think in our line of work we rely too much on metabolic rate testing. In fact, people are very insensitive to changes in their metabolic rate, so we are developing exoskeletons to accomplish something that people cannot really perceive.”
In contrast, the study authors wrote that user preference-based methods not only look at what users can perceive, but also allow them to prioritize qualities they deem important.
Subscribe to our sprawling newsletter
“Hardcore beer fanatic. Falls down a lot. Professional coffee fan. Music ninja.”