Application of an Ensemble Method to UAV Power Modeling for Cellular Communications

In this letter, we apply ensemble learning methods for the prediction of the ground (cellular base station) to air (flying node) received signal strength (RSS) at different heights, for future mobile communications. We model the RSS using different ensemble methods. Moreover, we propose a new ensemble method that combines results from five different methods. The proposed method also uses a recently introduced evolutionary algorithm, the Salp Swarm Algorithm, for weight optimization. The proposed method outperforms all the other methods and common ensemble methods. In this context, the produced results are compared to measurements using representative performance indices and exhibit satisfactory accuracy.