AI Smart Running Shoe based on nRF52832

Client

A sports biomechanics company creating a sensor-insole system for runners to analyze form and prevent common injuries like shin splints and plantar fasciitis.

Challenge

Building a flexible, washable insole with 16 pressure sensing points and a 6-axis IMU that maintains accuracy after thousands of foot strikes.

Product Concept & Architecture

Each insole contains a 16-point Force Sensing Resistor (FSR) array, an LSM6DSO32 accelerometer/gyroscope, and a Nordic nRF52832 MCU. The flexible PCB is encapsulated in TPU and sits between the insole layers.

Engineering Solution

The nRF52832 runs a real-time gait analysis pipeline: FSR data → center-of-pressure trajectory → foot strike classification (heel/midfoot/forefoot). The IMU provides cadence and ground contact time. A TinyML model on the device detects gait asymmetry patterns correlated with injury risk. Real-time audio coaching cues (via BLE to earbuds) correct form deviations mid-run.

Result

Validated against force plate reference (r²=0.96). 87% injury risk prediction accuracy in collegiate runner study.