<![CDATA[Burned fossil fuels from machines, especially vehicles, will release toxic gases which are harmful to human, even these toxic gases are colorless and odorless which makes them hard to be recognized by humans. Therefore, a research of detection system on toxic gas pollution caused by burned fuels is needed, by utilizing Occupational Safety and Health Association (OSHA) as a main reference to determine and classify the result of gas detection as a safe or dangerous condition according to their PPM level. There are 3 types of toxic gas detected in the system, first is Hydrocarbon (HC) gas detection using MQ-2 sensor, then Carbon Monoxide (CO) gas using MQ-7 sensor, and the last is detection of Nitrogen Oxide (NOx) gas with the MQ-135 sensor. The system is based on an ESP32 microcontroller using Support Vector Machine (SVM) as the classification method to classify the detection results from the MQ gas sensor. There are 60 training data which are taken from MQ gas sensor detection, and divided into 2 types of data, the first 30 data classified as safe condition, and the next 30 data classified as dangerous condition. The performance and accuracy test of the system uses 30 testing data which are also divided into 2 types, the first 15 data taken from a safe condition, and the next 15 data are taken on a dangerous condition. The detection and classification results, as an output, will be displayed on the LCD screen followed with a buzzer alarm sound if the SVM result is on a dangerous class. The accuracy obtained from the SVM classification test is 87%.]]>
Copyrights © 2022
