Random Forests Based Path Loss Prediction in Mobile Communication Systems

• Published in: 2020 IEEE 6th International Conference on Computer and Communications (ICCC) pp. 1246 - 1250
• Main Authors: He, Rongrong, Gong, Yuping, Bai, Wei, Li, Yangyang, Wang, Ximing
• Format: Conference Proceeding
• Language: English
• Published: IEEE 11.12.2020
• Subjects: Computational modeling; Data models; feature selection; Loss measurement; network planning; path loss prediction; Predictive models; Propagation losses; random forests; Training; Wireless communication
• DOI: 10.1109/ICCC51575.2020.9344905

When deploying communication systems, an accurate wireless propagation model is important to ensure the quality of service covering the region. Due to the complex radio environment, the traditional wireless propagation models need massive data for correction and calculation. To address this issue, this paper proposes a wireless propagation method to predict path loss. We use the random forest network structure to fit the complex model, accurately predicting the received signal power in the target area. To improve the training efficiency of the model, we construct the preliminary features according to the previous knowledge. A filtering feature selection method is adopted to select features as input of model. Evaluating the model on four typical terrains, the experiment results show that the proposed model outperforms the four existing models in all types of terrains.