Researchers in India have developed a wearable motion-sensing system that combines highly stretchable sensors with machine learning to track human movement accurately while consuming very little power. Designed to fit over ordinary clothing, the prototype could eventually support rehabilitation, sports training, and human–robot interaction, tells IEEE Spectrum.
The system addresses limitations of existing motion-tracking technologies, which can require substantial power or restrict natural movement. Researchers instead use the deformation of clothing as part of the sensing process. Six flexible piezoresistive sensors are positioned over the hips, knees, and ankles. As the wearer moves, the sensors stretch and convert physical deformation into electrical signals.
A machine learning algorithm running on a microcontroller analyzes synchronized signals from all six sensors. Rather than evaluating individual joints separately, the model examines short time windows of combined sensor data to recognize overall movement patterns.
Researchers tested the system with 12 participants performing activities that included walking on level ground and inclines and ascending or descending stairs. After training for individual users, the system classified movements with 99.83% accuracy. For a person who had not previously worn the system, accuracy reached 89%, an important result because body dimensions, movement patterns, and sensor positioning vary among users.
The sensing layer also requires only 0.318 milliwatts of power, allowing it to operate with a small battery. The flexible sensors demonstrated durability by surviving more than 12,000 cycles while being stretched to three times their original length.
The current prototype classifies movements, but researchers envision broader applications. Patients could monitor rehabilitation progress, while athletes could analyze movement during training. Wearable sensing could also help robotic systems respond dynamically to human motion rather than follow predetermined commands. The team now plans to test a larger, more diverse group and improve the system’s ability to accommodate individual movement differences under natural conditions.
