Researchers at Pohang University of Science & Technology in South Korea have developed a durable strain sensor that can detect complex body movements. The technology will be useful for patients undergoing physical rehabilitation, allowing physical therapists to assess their movements in significant detail and measure progress. Conventional strain sensors are often affected by heat and humidity, making them less durable as a wearable, and they typically detect only biaxial strain, providing less detail than these new sensors. The new technology uses computer vision, whereby an algorithm reviews digital images of the sensor deformation and calculates the movements of the sensor wearer.
Physical rehabilitation allows patients to regain mobility after an injury, disease or medical procedure. Physical therapists and clinicians are interested in characterizing the movements of such patients, such as determining their range of motion, gait, etc. This not only allows them to quantify a patient’s mobility, but monitor progress over time.
To date, researchers have developed plenty of new technologies that can help in this arena, from motion sensors to wearables. Strain sensors attached to the skin are a useful way to assess movement in specific regions of the body, but existing sensors have some limitations. This includes a complicated manufacturing process, and vulnerability to temperature and humidity, which is a drawback for an object expected to reside in close contact with the skin. They can also only typically measure strain in two axes.
To develop a better sensor, these researchers turned to optical sensors that employ computer vision to measure strain. This involves a miniature camera within the sensor that views a micropatterned silicon film that moves with the skin. The camera can view the film as it is deformed by the wearer’s physical movements, and then the system can use computer vision techniques to interpret these movements.
The researchers have called their technology computer vision-based optical strain (CVOS) sensors, and so far have shown that the sensors can detect rotational movements in three axes and can provide multiaxial strain mapping in real time. The system also incorporates an AI algorithm that works to reduce artifacts and errors in the data, and the sensors are robust, maintaining their performance over 10,000 cycles of use.
“The CVOS sensors excel in distinguishing body movements across diverse direction and angles, thereby optimizing effective rehabilitative interventions,” said Sung-Min Park, a researcher involved in the study. “By tailoring design indicators and algorithms to align with specific objectives, CVOS sensors have boundless potential for applications spanning industries.”
Study in journal npj Flexible Electronics: Real-time multiaxial strain mapping using computer vision integrated optical sensors