Joint Renewable Energy and Traffic Prediction for Cooperative Green Cellular Networks

Powering the base stations with renewable energy is a promising solution for green cellular networks. However, due to the double space-time instability of renewable energy and cellular traffic, there is a big gap between energy supply and traffic load. Developing combined dynamic state prediction for renewable energy and traffic is a viable method for the 'energy-traffic mismatches' problem, which can achieve effect network-wide resource utilization. But considering the high nonlinearity and numerical stability issues, it is challenging to obtain the prediction results accurately. In this work, the cubature Kalman filter is employed to predict the harvested renewable energy and cellular traffic jointly. Based on previous data, each base station predicts the renewable energy and traffic of the upcoming time slot. According to the prediction results, different base stations exchange renewable energy cooperatively. Experiment results indicate that the proposed joint renewable energy and traffic prediction-based cooperative algorithm can reduce the on-grid energy consumption.