Smart Control of Air Conditioning System Based on Number and Activity Level of Persons

Air conditioner (AC) is a device that requires high power. To reduce the energy waste by AC, the temperature setting should be adjusted to the room condition. It can be affected by the number of persons and the level of activity in the room. To save electricity consumption, this research is conducted to design an intelligent and automatic AC system based on identification of number of persons and activity level to control working temperature of AC using fuzzy logic refers to thermal comfort standard in Indonesia issued by SNI council. The system used USB camera as an image capture device. In image processing, Histogram of Oriented Gradient (HOG) method is used to identify the number of people, while background subtraction method is used to identify the activity level. This system is implemented into Raspberry Pi 3 as a single board computing. Test results show that the system is capable to detect people from a distance of 3m to 9m and distinguish 2 people within a distance of 30 cm to 150 cm, The system can also differentiate small, medium and high of activity levels. The test results show that the working temperature controlled by the designed fuzzy logic has the lowest value of 21°C and the highest value of 27°C which is in accordance with the thermal comfort standard that has been defined by the SNI council. The differences between working temperature and measured room temperature are within only 0.2°C to 1.2°C.