A non-stationary channel prediction method for UAV communication network with error compensation
Qiuyun Zhang,Tingting Yang,Chun Wu,FanRong Shi,Hong Jiang,Qiumei Guo,Liping Deng,Ying Luo
DOI: https://doi.org/10.1016/j.engappai.2023.106206
IF: 8
2023-04-07
Engineering Applications of Artificial Intelligence
Abstract:In an unmanned aerial vehicle (UAV) communication network , especially for mission-critical applications, ultra-reliable and low-latency communication (URLLC) of the control links has essential implications for realizing collision avoidance and real-time control of UAVs. However, the UAVs' mobility and the channel environment's variability may make the wireless channels highly non-stationary. Real-time and accurate Channel Status Information (CSI) acquisition is a critical challenge to implementing URLLC for UAV control links. Predicting the characteristics of non-stationary channels is profitable for formulating communication strategies to mitigate the effects of future channel fading. Nevertheless, the difficulty lies in extracting accurate prediction models for non-stationary channels to obtain accurate CSI. This paper proposes a more precise prediction by utilizing the CSI obtained as feedback from the receiver and introducing error correction. The method contains three stages, CSI collection and processing, channel tracking, and error correction. First, the transmitter collects the CSI feedback from the receiver and converts it to a stationary series. Secondly, the recent historical CSIs are used to track the time evolution of the channel based on an autoregressive integrated moving average (ARIMA) model. Next, the Gaussian process regression (GPR) model is used to establish the prediction error regression model and obtain a more accurate prediction. Finally, the effectiveness of the proposed method is verified based on a UAV wireless communication network application scenario in an urban environment. Simulation results show that the proposed method outperforms other methods regarding RMSE and reliability in high-dynamic UAV communication scenarios. Significantly when the channel changes rapidly, this method can respond to the changes faster and predict the CSI more accurately.
automation & control systems,computer science, artificial intelligence,engineering, electrical & electronic, multidisciplinary