A Smart City IoT Framework with BiLSTM-based Rainfall Prediction and Microservice Architecture

Internet of Things (IoT) has experienced significant growth, impacting various aspects of society. However, the efficient management of IoT devices, particularly in smart cities, is still challenging due to the diverse connection types and their massive collected data. Therefore, in this paper, we introduce an IoT device management system that monitors and configures all connected devices through the HTTP and MQTT connection protocols. We also build a rainfall prediction model based on a BiLSTM architecture and incorporate it into the proposed system using the microservice architecture. The experimental results on two experimental datasets show that our proposed model achieves the highest F1-score compared to other approaches, which shows the potential of the proposed system for predicting rainfall in real-life scenarios.