Initial Study of Building Smart Air Pollution Sensors with the Decision Tree Algorithm

Abstract

In the last decade, Indonesia shifted from one of the cleanest countries in the world to one of the countries with high air pollution, especially in big cities like Jakarta, Surabaya, etc. This paper presents a preliminary study of the building of a smart sensor node system to monitor air quality for urban environments. Sampled air parameters were CO, CO2, and CH4. Incoming sensor data were classified using rules from the calculation of the entropy and information gain done before in the Decision Tree Algorithm rule finding steps. The classification is done on a Raspberry Pi device using conditional if-else on the programs. The if-else condition makes the program could achieve air quality results. The decision results in the form of warning alert or data reports packed on a Raspberry Pi device. By then using LoRa transceiver, a medium-range communication system, these results sent to a PC server. Reliability testing has been carried out starting from the sensors, algorithm processing capability, until the transmission process with Time Division Multiplexing technique. Data is sent alternately between one node to another. Based on the testing results, the CO sensor shows good performance at pollutant levels up to 300 ppm, so do the CO2 sensor at values up to 500 ppm, and the CH4 sensor at values up to 250 ppm. Test measurements have been carried out in outdoor environments for feasibility of the transmission system that has designed.