Smart Shirt Tracks Workouts, Washes Easily

Researchers at the University of Cornell “T -Shirt” developed a highly accurate exercise and position data without compromising the appearance, comfort or washing capacity. Unlike many existing smart clothes, the shirt – the volatile – does not depend on huge sensors or narrow and restricted suitability. The shirt has possible applications across various fields, including medicine and athletics.

Seamfit uses flexible connected threads thread in the neck, arm and side cells for a standard cotton shirt otherwise. He says, “When the body of its wearing changes, the poles of the seam, that is, the connected strings, the distortion and change of its own curtains, the conjunction between the body of its wearer, a large conductor, and the seams change.” Catherine UPhD student in the field of information science in Cornell. “By monitoring the capacity changes, we can conclude the body that puts/movements that stimulate these changes.”

A small -semester circle plate in the neck line is recorded and sent via Bluetooth to a laptop. There, the artificial intelligence model analyzes the signal to identify exercises and the number of repetitions.

The researchers tested the performance of the shirt in 15 volunteers and described the results in a paper Posted in March in ACM facts on interactive technologies and mobile phone, can be worn and semi -place everywhere.

High exercise detection engineering

In the study, each volunteer conducted 14 different exercises, including stab wounds, seating and header. Without the user’s calibration, Seamfit showed the accuracy of the discovery of the exercise by 89 percent, the accuracy of the exercise classification of 93.4 percent, and was suspended by less than a average delegate when the actors’ account is calculated within a group-based group, washing, and appropriate.

Short -sleeved seamfit design means some joints, such as elbows and wrists, not covered directly by electrodes. However, their movements can be tracked indirectly, says Yu. Seamfit can capture movements from other parts of the body-while isolated leg movements generate weak signs of the shirt, most lower body exercises include a higher body movement to produce consistent patterns that can be tracked.

Seamfit still requires some preparation before it goes to washing. The researchers merged a detachable circuit board that must be removed before throwing the shirt in a washing machine. However, the strands connected to the clothes keep the dress, keeping its conductivity through multiple laundry cycles – although Yu says that its performance is slightly decomposing over time, which provides differences in the quality of the sign after washing enough. To address this, the team trained the data processing pipeline on data collections collected after multiple washes, allowing the system to adapt to these changes and maintain reliable accuracy.

The researchers used the connected thread in the neck, arms and side layers for a standard cotton shirt to track the movements of its wearer.Louis Deeptro/Cornell University

“Practical design options such as removable circuit board and the machine washing the machine directly treat user’s common concerns about maintenance and comfort,” he says. Rong YinAssistant Professor in the Department of Textile, Chemistry and Science Engineering at North Carolina State University, who did not participate in the research.“This approach centered around the user significantly increases the possibility of widespread adoption, not only among fitness lovers but also in clinical environments for physiotherapy and rehabilitation.”

Beyond fitness tracking, seamfit can have applications in robots. Because it provides an unobtrusive and comfortable way to collect data over the extended periods, Seamfit can help train human robots to simulate human movements better, says Yu. By capturing data on human movement, the shirt can also improve the integrity of human bobote reactions. YIN adds that seamfit can also have applications in human interaction and shapes “where understanding of micro human movements is extremely important.”

To understand the signals coming from the smart shirt, the researchers developed a three -portal AI system. A lightweight random forest model discovers whether its wearer is practicing or rest. Deep encryption and rejection model determines what is being implemented. And the algorithm of the extension peak detector, which does not require any training, calculates the number of repetitions by its wearer.

The researchers used the “participating leave” approach to test the extent to which their artificial intelligence models were generalized to new users. This included training forms on data from all participants except for one, and then tested on the data from the participant who was excluded. By repeating this process for each individual in the study, the team confirmed that the system works even with different people, and the suitability of a different shirt, and the signal changes from the washing.

One day, wearing an ordinary smart shirt may be like wearing cotton today-and according to Yu and its colleagues, Seamfit represents a step towards that future.