THE EFFECTIVENESS OF MULTI-SEN SOR SYSTEMS IN MONITORING POLLUTION LEVELS

摘要: A multi-sensor network technology is proposed in this study. This technology is applied at the semi-closed motorcycle parking area of Politeknik Negeri Sriwijaya. The sensor node is composed of a multi-sensor, and there are six characteristics of pollutant levels that are continuously measured. These parameters include Carbon Monoxide (CO), Carbon Dioxide (CO2), HydroCarbon (HC), temperature, humidity, and particulates (PM10). The data can be monitored using a mobile phone. A Support Vector Machine (SVM) is supplied to the system to enhance the classification validity and accuracy. The proposed device gives 95.02% accuracy and only 4.98% classification error for node 1, while for node 2, it provides 99.3% accuracy and 0.05% error, and for node 3, it gives 95.03% accuracy and 4.97% error. The resulting device is efficient for economic sensor-based air quality monitoring.