SmartSense:AI-Driven Low-Cost Wearable for Real-Time Symptom Detection and Risk Estimation

Sammanfattning

Asthma is a chronic respiratory condition that demands continuous symptom monitoringto prevent severe health complications. This project aims to create Smart-Sense, a low-cost, wearable prototype designed to detect early symptoms of asthmausing real-time physiological monitoring. The device integrates a conductive rubberband to track chest expansion and a digital microphone for cough detection. Amicrocontroller samples the sensor data, performs onboard signal processing usingFFT for breathing analysis, and applies a quantized neural network model for AIbasedcough classification. Processed data is transmitted to a companion mobileapplication via BLE, enabling real-time visualization, alerts, and symptom tracking.Emphasis is placed on local data processing and energy efficiency to supportextended autonomous operation. Testing showed promising results in detectingrespiratory patterns and classifying cough events, although challenges such as backgroundnoise sensitivity and sensor nonlinearity remain. SmartSense demonstratesthe feasibility of affordable, embedded AI in wearable health technologies and laysthe groundwork for future clinical validation and commercial development.