TY  - BOOK
AU  - LOW JIA CHENG (TP038654)
AU  - Dr. Tan Chye Cheah
TI  - A COMPUTER VISION-BASED APPROACH ON ENERGY SAVINGS:  AUTOMATED SMART HVAC CONTROL
AV  - PG-21-0139 
PY  - 2019///
CY  - Kuala Lumpur
PB  - Asia Pacific University
KW  - Energy efficiency
KW  - Computational intelligence
KW  - Automation
KW  - Application software
N1  - A project submitted in partial fulfillment of the requirement for the degree of B.Sc (Hons) in Computer Science (UC3F1902CS)
N2  - HVAC system in premises are often left running unattended for long hours even during lunch or after school or work hours which resulted in electricity wastage and high electricity bills. On average, HVAC systems are the second most electricity consuming appliances, reaching up to 28% of electricity use in United States (Simpkins, 2010). Although sensors designed for such use cases have existed, most of the sensors operate without the ability to adjust air conditioner in response to the number of occupants. Presence sensor can detect human presence, but it can only turn air conditioner on or off. As such, this project proposes a solution by integrate computer vision to dynamically adjust air conditioner. Using deep learning model to recognise person and chair, the system calculates occupancy rate which is then used to determine the fan speed. The chairs count is used in the system to adapt to classroom size since there should be more seats in bigger classroom and vice versa. The proposed system then integrates to existing HVAC systems by simply sending infrared instructions using ESP8266 Infrared Transmitter. With Azure IoT Hub’s Device Method, Raspberry Pi transmits specific instruction to ESP8266 remotely. An Android application was also built to display current status and live camera feed. The statuses are sent in real time after each frame is processed through IoT Hub’s Cloud-to-Device messaging feature
ER  -