Intelligent HVAC Control System

Abstract

Comfortability where occupant is presence is the subject of marketing in many sectors. While there are many areas that contribute to comfortability, 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. State of the art shows many control approaches captured/compared and provide great potential but also agree that there is still room for improvement. In addition, other reviews were also compared to examine their studies in this area. Some of the earlier approaches use standard control features but as time progress and more tools and technology become available, the HVAC control development progressed even further to integrate artificial intelligence and machine learning and open new opportunities for improvement/optimization. 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 provide results as new combined approach for HVAC control. Prediction algorithm LDA shows approximately 79% accuracy score for prediction which scores 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. Future work for additional development/improvement are also mentioned in the conclusion in future work section.