Overview
INVISION is a wearable armband that combines optical, acoustic, electrical, and temperature sensors to provide comprehensive vital signs monitoring for under $10.
Designed for at-risk individuals such as the elderly and those with chronic conditions, INVISION enables continuous health monitoring, early anomaly detection, and remote caregiver alerts.
Key Characteristics
The Problem
This project was inspired by personal experience with elderly family members needing continuous health monitoring.
Market Gap
Fitness Wearables
Track basic vitals (heart rate, steps) but limited functionality, not designed for early anomaly detection, fitness-focused only.
Clinical Equipment
Monitor critical vitals but high cost, mobility issues, continuous power needs, requires expert interpretation.
Technical Implementation
INVISION uses an ESP32 microcontroller connected to four sensor modules, powered by a 3.3V battery. Data flows to cloud storage for processing, a mobile app for the user interface, and an alert system for notifications.
How It Works
Uses the Beer-Lambert Law to analyze absorbance across different wavelengths of light as they pass through the skin. Different wavelengths are absorbed differently by oxygenated vs deoxygenated blood.
Key Measurements
- Pulse rate from green light absorption patterns
- Blood oxygen saturation (SpO2) from red/infrared ratio
- Potential glucose estimation from NIR wavelengths
Components
Multi-wavelength LED array (green, red, infrared) paired with photodiode detector.
How It Works
A thermistor changes resistance based on temperature. Using a voltage divider circuit with a known resistor and input voltage, we measure the thermistor voltage and calculate temperature using datasheet calibration values.
Key Measurements
- Skin temperature (contact sensor)
- Ambient temperature (reference sensor)
- Temperature trends over time
Clinical Relevance
Elevated skin temperature can indicate fever, infection, or inflammation. Sudden changes may signal health anomalies requiring attention.
How It Works
The ESP32 drives a small speaker to generate discrete sensing signals. A microphone then captures vibrations transmitted through the skin, detecting the characteristic "lub-dub" heart sounds (S1 and S2).
Key Measurements
- Heart rate from sound interval timing
- Heart sound patterns (S1, S2 detection)
- Potential murmur or arrhythmia detection
Signal Processing
Digital filtering isolates heart sounds from ambient noise. Pattern recognition identifies abnormal cardiac rhythms.
How It Works
Multiple closely-spaced electrodes on the armband detect the electrical signals generated by the heart. The configuration allows 4-lead EKG-equivalent measurements from a single wearable location.
Key Measurements
- Heart rate and rhythm
- QRS complex detection
- Arrhythmia indicators
- Heart rate variability (HRV)
Validation
Single lead signals validated consistently against clinical-grade equipment. Electrode placement optimized for reliable skin contact.
System Architecture
ESP32 Microcontroller → Four Sensor Modules → Cloud Processing → Mobile App + Alert System → Healthcare Provider
Research Questions
Can a low-cost, multi-modal biosensor system detect physiological deviations from baseline using optical, acoustic, and electrical sensors?
How can a wearable multi-modal biosensor suite effectively detect early vital sign anomalies to improve health outcomes?
What is the optimal method for integrating multiple sensing modalities in a compact, wearable design?
How can real-time signal processing be achieved with minimal power consumption?
How can AI enhance real-time health monitoring and early intervention?
Applications
Remote Home Monitoring
For elderly, immunocompromised individuals, and people with disabilities
Chronic Disease Management
Continuous monitoring for diabetes, heart disease, and other conditions
Post-Surgery Recovery
Track vital signs during recovery at home
Rural Healthcare
Remote patient monitoring for underserved communities
Workplace Safety
Monitoring for high-risk occupations
Fall Detection
Emergency detection and automatic alerts
Future Work
Multi-channel Alerting
Develop multiple forms of alerting users and healthcare providers
Improved Form Factor
Make the device less obtrusive using flexible printed circuit boards (PCB)
Health App Integration
Connect with Apple Health, Google Health, and other platforms
Intelligent Data Collection
Optimize battery life with smart sampling algorithms
ML-Based Predictions
User clustering and predictive models for early intervention
Team
Adi Desai
Co-developer
Anika Shah
Co-developer