Intelligent Internet of Things Using Artificial Neural Networks and Kalman Filters for Energy Management Systems

Abstract

Internet of Things (IoT) concepts are widely used for controlling and managing electrical power, especially for residential and commercial buildings. However, these controls are still condition-based methods that are limited in decision-making and inflexible operation. In addition, the transmission of data from sensors over the internet may be interrupted or scrambled, resulting in a controller processing error. This paper proposes an artificial intelligence (AI)-based approach for controlling IoT devices to enhance the ability of the controller to operate intelligently. A neural network technique is used to optimize the controller operation in the IoT system. The state estimation approach using the Kalman filter (KF) algorithm is proposed to reduce data errors and increase the reliability of the IoT control system. The proposed intelligent IoT approach is implemented for energy management and tested on a laboratory case study to minimize energy use for the lighting system. The experimental results show that the proposed method decreases 49.56% of electricity consumption and reduces the data variance from sensors by 77.13% compared to the conventional system without intelligent control. The test results indicate that integrating AI and KF with the IoT system can efficiently and effectively control and manage the lighting system.