Beyond 5G for digital twins of UAVs

Abstract

The purpose is to explore the application effects and limitations of Unmanned Aerial Vehicle (UAV) in 5G/B5G (Beyond 5G) mobile and wireless communication. Based on 5Gcommunication, the deep learning (DL) algorithm is introduced to construct the UAV Digital Twins (DTs) communication channel model based on DL. The Coordinated Multi-point Transmission (COMP) technology is adopted to study the interference suppression of UAVs. The key algorithm in the physical layer security is employed to ensure information communication security. Finally, the model constructed is simulated and analyzed. The transmission error rates and transmission estimation accuracy of several algorithms, including the proposed algorithm and ordinary Deep Neural Networks (DNNs), are compared under different Signal-to-Noise Ratios (SNRs). Results find that the convergence speed and convergence effect of the proposed algorithm has prominent advantages, presenting strong robustness; the proposed algorithm's estimation accuracy is about 150 times higher than the traditional algorithms. Further analysis reveals that the proposed algorithm's accuracy reaches 82.39%, which increases by at least 3.2% than other classic machine algorithms. The indicators of Precision, Recall, and F1 are compared as well. Apparently, the Precision, Recall, and F1 values of the proposed algorithm are the highest, while the transmission delay is the smallest. Therefore, the constructed UAV DTs wireless communication channel model has strong robustness and further reduces UAV limitations, providing a reference for improving UAV system performance in the later stage.