Realtime Monitoring System Towards Waste Generation Management

Waste management is a significant concern to protect the environment and the population’s health. However, due to its uncertainty and high variability, waste generation has dynamic nurture that results in ineffective, inaccurate, and unreliable waste management. Therefore, the existence of the IoT, information systems, and artificial intelligence can help support sustainability and decrease the amount of waste. This research proposes a monitoring system equipped with a forecasting feature and implemented in a mobile application to overcome the limitations of conventional waste management systems. This system is built on two Internet of Things (IoT) architecture nodes with an android-based mobile application interface. Measures the unfilled level of the bin, processes it, and sends it to the database. The data was then computed using Levenberg Marquardt’s Artificial Neural Network to predict the height of the garbage. The results of the IoT communication show that the average delay in sending data to the database is 2. 886s and 2. 912s, with 0% packet loss. The correlation coefficients generated in the Levenberg Marquardt model training process are 0.925 and 0.965. In addition to displaying garbage height data and prediction results, this system application can also display the location of the bin and receive notifications. This monitoring system has also been tested directly on the Undip waste manager using the SUS questionnaire. Based on these tests, the SUS score of70.2 showed that the application already had good usability.