Abstract

The convergence of Artificial Intelligence (AI) with Internet of Things (IoT) technologies, often referred to as AIoT, is transforming aspects of modern life, such as smart cities. This transformation, however, brings with it challenges, including energy management. In addressing this issue while upholding responsible AI principles, it is important to prioritize the sustainability of AIoT solutions by a promising approach which is using renewable energy sources. While renewable energy offers numerous advantages, its intermittent nature necessitates effective power management systems. Developing a power management system serving as a decision-making platform for AIoT-driven solutions is the goal of this study. This platform contains two critical components: accurate forecasts of battery "State of Charge" (SoC), and the implementation of appropriate control strategies. These strategies include adjusting energy consumption patterns to ensure stable and reliable system operation. This study focuses on accurate battery SoC forecasting, to this end, an experiment has been designed, and a data logging system has been developed to produce suitable data since publicly available datasets do not align with the specific characteristics and requirements of the research. The SoC forecasting in this study has been addressed as a multivariate and multi-step time series forecasting problem, where various machine learning and deep learning models including Decision Tree (DT), Random Forest (RF), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Bidirectional Long Short-Term Memory (Bi-LSTM), and Bidirectional Gated Recurrent Unit (Bi-GRU) were benchmarked. Extensive evaluations have been conducted for different forecasting horizons on datasets with varying time intervals. It is concluded that the Bi-GRU model outperformed other models across datasets with varying time intervals and forecast horizons according to Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) evaluation metrics.