[en] Satellite and unmanned aerial vehicle (UAV) net-works have been introduced as enhanced approach to providedynamic control, massive connections and global coverage forfuture wireless communication systems. This paper considersa coordinated satellite-UAV communication system, where theUAV performs the environmental reconnaissance task with theassistance of satellite in a hostile jamming environment. To fulfillthis task, the UAV needs to realize autonomous trajectory controland upload the collected data to the satellite. With the aid ofthe uploading data, the satellite builds the environment situationmap integrating the beam quality, jamming status, and trafficdistribution. Accordingly, we propose a closed-loop anti-jammingdynamic trajectory optimization approach, which is divided intothree stages. Firstly, a coarse trajectory planning is made accord-ing to the limited prior information and preset points. Secondly,the flight control between two adjacent preset points is formulatedas a Markov decision process, and reinforcement learning (RL)based automatic flying control algorithms are proposed to explorethe unknown hostile environment and realize autonomous andprecise trajectory control. Thirdly, based on the collected dataduring the UAV’s flight, the satellite utilizes an environmentsituation estimating algorithm to build an environment situationmap, which is used to reselect the preset points for the first stageand provide better initialization for the RL process in the secondstage. Simulation results verify the validity and superiority of theproposed approach.
[University of Luxembourg > Interdisciplinary Centre for Security, Reliability and Trust (SNT) > SigCom >]
