Overview
- Rutgers engineers published a Science Advances paper describing a prototype shoe that generates electricity from each step and runs a tiny AI model inside the shoe to classify walking, running, and stair climbing.
- The sole contains a triboelectric generator that turns pressure and friction into bursts of charge and a power-management circuit that the team says makes up to 120 times more usable energy than a conventional method.
- The compressed edge-AI model uses an accelerometer and fits the shoe’s tiny processor, consuming about 86 microwatts and achieving 95.4% accuracy on 15-second motion segments in lab tests.
- Tests so far used data from four healthy volunteers ages 23–26, so the device is an early proof-of-concept, not a medical product, and the inventors have filed a Rutgers-assigned U.S. provisional patent.
- If scaled and clinically validated, the approach could let people wear ‘worry-free’ shoes that monitor gait over days to track fall risk, disease progression, or rehabilitation without frequent recharging.