Researchers from the College of Oxford, College School London, and Xi’an Jiaotong College have developed an “AI-enabled” piezoelectric wearable that, they are saying, represents a breakthrough in joint well being monitoring — passing its knowledge by an tiny machine studying (tinyML) mannequin operating on a microcontroller growth board.
“Joint well being is important for musculoskeletal (MSK) circumstances which are affecting roughly one-third of the worldwide inhabitants,” the researchers say of the main target of their work. “Monitoring of joint torque can provide an essential pathway for the analysis of joint well being and guided intervention. Nonetheless, there isn’t a expertise that may present the precision, effectiveness, low-resource setting, and long-term wearability to concurrently obtain each fast and correct joint torque measurement to allow danger evaluation of joint harm and long-term monitoring of joint rehabilitation in wider environments.”
A comfortable, versatile, self-powered sensor may assist present long-term joint well being monitoring — due to a machine studying mannequin operating on a low-cost microcontroller. (📷: Chang et al)
That’s, there was no expertise with addressed all of those necessities — till the researchers developed it. The crew’s system is a versatile, comfortable, and light-weight wearable torque sensor based mostly on boron nitride nanotubes (BNNTs) on a polydimethylsiloxane (PDMS) substrate, which makes use of the peizoelectric impact to each seize knowledge in regards to the wearer’s knee actions and harvest its personal energy.
The key to the system’s success: a tiny machine studying (tinyML) mannequin, educated on knowledge captured from an Arduino Nano 33 BLE operating and deployed to an STMicroelectronics STM32 NUCLEO F401RE growth board through TensorFlow Lite and X-Dice-AI, which processes the sensor’s piezoelectric outputs and maps them to bodily traits together with torque, angle, and loading — offering steady monitoring as an alternative of present point-in-time measurement approaches, with out inflicting discomfort to the consumer.
The wearable (b) makes use of the piezoelectric impact to each energy itself and measure the knee’s motion, feeding knowledge to a tinyML mannequin operating on a microcontroller. (📷: Chang et al)
“The proposed system has the potential to advance international efforts in joint well being monitoring, the administration of MSK circumstances, rehabilitation, ageing issues, and broader purposes in private healthcare,” the crew claims. “Future analysis will prioritize enhancing the system’s adaptability, scalability, and inclusivity.”
The researcher’s work has been printed within the journal Nano-Micro Letters.
