Overview:
This project is currently in development. I'm exploring ways to detect early signs of plantar fasciitis in gymnasts by tracking foot movement with an accelerometer. The goal is to capture subtle, impact-related changes over time—offering a noninvasive method to monitor stress before it becomes an injury.
Materials and Procedure:
Hardware Platform: Built on the Feather ESP32-S3 (no PSRAM), chosen for its compact form and low power consumption—ideal for lightweight, wearable applications.
Motion Sensing: Movement is captured using the Adafruit ICM-20649 accelerometer, selected for its wide ±30g range and high 1125 Hz sampling rate. This enables accurate tracking of impact forces, centripetal acceleration, and lateral movement during high-intensity routines.
Power Supply: A small, rechargeable battery powers the system, keeping the overall form factor minimal and unobtrusive.
Data Handling: The Bluefruit Connect app enables real-time data collection and processing, with the option to store sessions for post-practice analysis.
Attachment Method: The sensor is secured using a flexible strap and athletic tape, allowing for stable placement around the ankle without restricting movement or performance.
Key Design Components:
Responsiveness and Real-Time Detection: The system is designed to capture and process foot movements instantly—focusing on identifying sudden shifts or irregular patterns that may signal overuse or early-stage injury.
Precision Monitoring: Fine motor data is critical. The sensor must detect subtle variations in gait and pressure distribution that could point to the gradual onset of plantar fasciitis—before it escalates into a performance-limiting issue.
Challenges and Problem Solving:
Sensor Calibration in Motion: One of the key challenges was ensuring accurate data collection during dynamic movement. Gymnastics involves rapid changes in force and direction, so calibrating the accelerometer to consistently capture meaningful data—without interference from noise or irrelevant movement—required fine-tuning sampling rates and sensitivity thresholds.
Real-Time Transmission & Data Integrity: Bluetooth transmission had to be both fast and stable to enable real-time plotting without data loss. Iterative code adjustments were made to optimize the balance between data resolution and transmission speed, particularly when operating at higher sampling rates up to 200Hz.
Wearability and Secure Attachment: The device needed to be lightweight, unobtrusive, and firmly secured to the gymnast’s ankle without restricting movement. Testing various strap and tape configurations helped identify a setup that ensured sensor stability while preserving full range of motion.
Motion Data Analysis: The intervals and rates are according to how many samples are collected in one second. For example, I can reduce the sleep time for higher sampling rates: 0.0025 seconds is about 40Hz, 0.01 seconds is about 100Hz, and 0.005 is about 200Hz.
Outcomes and Impact:
This early-stage prototype successfully transmitted real-time data wirelessly while capturing detailed acceleration profiles during gymnastics movements. By mapping these signals to specific types of force and directional stress, the system lays the groundwork for detecting biomechanical patterns associated with overuse injuries. Continued testing will refine the thresholds and improve predictive accuracy.
Reflections:
This project has deepened my understanding of motion sensing and wearable integration in athletic contexts. Working at the intersection of embedded systems and movement science, I’ve learned how subtle sensor positioning, data timing, and real-world constraints affect not only technical performance, but also the comfort and usability of a device. The process has strengthened my skills in data-driven design and reinforced my approach to building systems that respond intelligently to the body in motion.