Real-Time In-Vehicle Air Quality Monitoring System Using Machine Learning Prediction Algorithm

Area of Science:

• Environmental Science
• Automotive Engineering
• Data Science

Background:

• In-vehicle air quality significantly impacts driver health and safety.
• Real-time monitoring and prediction are crucial for proactive management of cabin air.
• Existing systems often lack predictive capabilities for future air quality conditions.

Purpose of the Study:

• To develop a real-time, cloud-based system for monitoring and predicting in-vehicle air quality.
• To utilize machine learning algorithms for analyzing sensor data and forecasting cabin air conditions.
• To assess the potential of air quality data in predicting driver drowsiness and fatigue.

Main Methods:

• Integration of five sensors (CO2, particulate matter, speed, temperature, humidity) into a cloud-based system.
• Real-time data collection and cloud storage from the vehicle cabin.
• Development and evaluation of machine learning models (Multilayer Perceptron, Support Vector Regression, Linear Regression) for air quality prediction.

Main Results:

• Support Vector Regression demonstrated superior performance in predicting in-vehicle air quality.
• The SVR model achieved a high coefficient of determination (R^2) of 0.9981, indicating excellent linearity.
• The system successfully correlated air quality parameters with potential driver fatigue indicators.

Conclusions:

• The developed cloud-based system provides accurate real-time and predictive in-vehicle air quality monitoring.
• Machine learning, particularly Support Vector Regression, is effective for forecasting cabin air conditions.
• This technology has the potential to enhance driver safety by monitoring air quality and predicting fatigue.