Experiments on Energy Optimization in Smart Residences

Abstract

Internet of Things (IoT) devices emerge to integrate devices (or "things") into the Internet, with limited computational resources, thus IoT networks have been integrated into the cloud and fog paradigms to mitigate these drawbacks. On the other hand, artificial intelligence (AI) techniques have shown to be efficient in several different areas, especially for data classification and prediction. In this paper, it is proposed a model of an RFID access control system with a neural network fog module that, considering the access data in a smart condominium with 300 homes, can estimate schedules and figure out when the homes are unoccupied, and by using the long sleep technique during that time, up to 9.9% of additional energy savings could be obtained. Future work can use this knowledge for developing a variety of optimizations and to improve the residents’ quality of life. The viability of this model is demonstrated by a fog network prototype.