Intelligent HVAC Control Prediction

Abstract

Comfort where an occupant is present is the subject of marketing in many sectors. This research paper focuses on heating, ventilation, and air conditioning (HVAC) in the transportation sector. A literature survey has been conducted to understand historic HVAC control and optimization approaches. Many control approaches were captured/compared, and this provides great potential, but also shows that there is still room for improvement. This research explores a unique control opportunity using linear discriminant analysis (LDA) to predict the occupant, and then follows it with kalman decomposition (KD) for real time controllability/observability post-LDA operation. Integrating these two tools provides results as new combined approach for HVAC control. Prediction algorithm LDA shows approximately 79% accuracy score for prediction, which is above average when compared to other algorithms and sensors used. KD is manipulated to be controllable and observable to maintain cabin temperature in real-time once the occupant is identified.