Researchers at Meta and the University of Twente have developed MyoSuite, a realistic musculoskeletal simulation tool that runs up to 4,000 times faster than state-of-the-art prosthetics.
The AI-powered digital models in MyoSuite can build realistic musculoskeletal simulations as many as 4,000 times faster than current prosthetics. The researchers claim that they can train the models to do certain complex motor movements, such as twirling a pen or rotating an arm.
Although the main goal of the tool is to model complex human movement to help develop prosthetics, it can also potentially be used for the design of new physical rehabilitation and surgery techniques.
MyoSuite model / Meta
“If we could predict the outcome of a robotic therapy beforehand, then we could optimise it for a patient and deliver a truly personalised and cost-effective treatment,” said Professor Massimo Sartori, one of the researchers in the team.
“We hope that diverse features supported by our framework will open new opportunities in understanding neuromechanical systems interacting with artificial robotic agents,” Sartori said.
Although this ML program is Meta specific, the company plans to make MyoSuite open source to support other researchers in this field. Meta is not the first company to make advances in the field of prosthetics, as Intel recently launched a ‘neuromorphic’ deep-learning chip that they said would make prosthetic limbs better.
When it comes to the metaverse, MyoSuite could also be integrated in programmes like Horizon Worlds, to develop more realistic avatars.
Source: E&T
