Intelligent Mat Predicts Posture Based on Tactile Movement
Wouldn’t it be great if someone could provide input on your yoga postures while you do them? An online instructor is a great option but sending somebody your camera feed over the internet for almost an hour just for this purpose seems overkill. Well, in the near future, your yoga mat could provide inputs on improving your postures. Well, that could be one of the many applications of this new technology being developed by researchers at MIT.
Like you, the researchers at the Computer Science and Artificial Intelligence Laboratory (CSAIL) at MIT are also worried about the excessive usage of video cameras in our personal lives. They set out a design a system that could estimate human poses without these cameras. Since we spend most of our time in contact with the ground, the team decided to capture the information we generate during this contact and use it to train a cameraless system.
Well, they had to begin with a camera to correlate the impact of human actions on the floor. The impact was measured using a specially designed tactile mat made out of a pressure-sensitive film, conductive threads, and over 9000 sensors. Each sensor converted the impact of human contact into an electric signal. By synchronizing the visual feed and the electric signal data, the researchers trained a neural network to identify the activity, a human was performing.
Although the number of sensors seem appalling at first, the mat costs less than $100 to manufacture per unit.
“You could envision using the carpet for workout purposes. Based solely on tactile information, it can recognize the activity, count the number of reps, and calculate the amount of burned calories,” said Yunzhu Li, a Ph.D. student a co-author of the paper.
Since pressure data was obtained largely from the lower body, the system is less accurate when predicting movements of the upper body. In the future, the researchers want to improve detection when two individuals are on the mat and estimate information pertaining to the height and weight of the individual.
The researchers are confident that their technology is a step toward improving self-powered personalized health care, smart homes, and gaming in the future.
