An Efficient Edge Analytical Model on Docker Containers for Automated Monitoring of Public Restrooms in India

An Efficient Edge Analytical Model on Docker Containers for Automated Monitoring of Public Restrooms in India

연구 분야: Software Development

학회: 2020 IEEE International Conference on Advanced Networks and Telecommunications Systems (ANTS)

The Government of India proposed “Clean India Mission” in 2014 that has led to the establishment of many public restrooms across the country. The purpose of this mission is to provide hygienic toilets to prevent open defecation and eradicate water-borne diseases. These restrooms are assigned human resources. However, the majority of them are in bad condition for usage. For increasing the availability of toilets, there is a requirement of a smart sanitation system for its automated monitoring. Moreover, rural and semi-urban restroom locations suffer from limited network connectivity. Thus, restroom state analysis needs to be performed near the site for seamless monitoring. This paper proposes an edge analytics framework based on the Internet of Things (IoT) and fog computing to analyze restroom data at the network edge and provide intelligent decisions. The machine learning model runs over Docker containers for analyzing data for the resource-constraint fog nodes. The choice of a machine learning model for the analysis depends on the model's performance in classification and on Docker containers over Raspberry Pi. Results show Adaptive Neuro-Fuzzy Inference System (ANFIS) is suitable over Docker containers for handling restroom data at the edge of the network.