Inspiration
Asthma affects over 300 million people worldwide, yet many struggle to tell when their breathing changes from normal to dangerous. During workouts or exposure to allergens, it’s often hard to know if shortness of breath is asthma, anxiety, or something else. We wanted to build a tool that helps users hear and detect the difference—empowering early intervention and peace of mind.
What it does
AsthmaOrNah listens to breathing sounds through a phone mic or ESP32-S3 Sense and analyzes them in real time. By measuring airflow rhythm, pitch stability, and harmonic distortion, it identifies patterns like wheezing or airway restriction that could indicate asthma or other respiratory problems (such as vocal cord dysfunction or tracheal collapse). The device gives immediate visual and audio feedback to help users understand their breathing.
How we built it
We processed live microphone input using Fast Fourier Transform (FFT) and band-energy ratios to detect abnormal sound frequencies typically associated with wheezing (100–1000 Hz range). The ESP32-S3 runs a lightweight C++ and MicroPython script to capture and filter the signal, while a Flask dashboard displays real-time graphs of breathing waveforms and spectral shifts—completely AI-free and transparent.
Challenges we ran into
Designing filters sensitive enough to detect subtle wheezes without false alarms. Calibrating the microphone gain and noise cancellation for consistent results. Distinguishing asthma-like sounds from environmental noise and coughs.
Accomplishments that we're proud of
Built a fully functional real-time breathing analyzer with no AI dependency. Visualized frequency and amplitude changes of real breathing samples. Demonstrated differences between normal breathing, asthma, wheezing, and EILO. Created an open, explainable tool that anyone can understand and replicate.
What we learned
We learned that classical signal processing still holds enormous power for health monitoring—especially when transparency and interpretability matter. Breathing patterns can tell powerful stories when you simply visualize them correctly.
What's next for asthmaornah
Add multi-sensor inputs (e.g., airflow, temperature, pulse oximetry) for richer context. Test with asthma patients to refine detection thresholds. Create a mobile companion app that records and visualizes daily breathing logs. Open-source the project to encourage more low-cost, AI-free health innovation.
Log in or sign up for Devpost to join the conversation.