Abstract:

The security and privacy protection of wireless communication networks constitutes a critical challenge in the evolution toward 6G space-air-ground network. To address diverse privacy threats in complex scenarios characterized by three-dimensional coverage and multi-domain collaboration， an uplink secure transmission scheme for space-air-ground network under dynamic adversarial environments is proposed. The system model encompasses four core nodes： a satellite， a legitimate Relay Unmanned Aerial Vehicle （R-UAV）， a malicious Eavesdropping UAV （E-UAV） and terrestrial users. The satellite is configured to receive relay signals solely from the R-UAV， simulating potential harsh channel conditions and the application requirements of cooperative communications. The E-UAV cruises within the communication airspace to opportunistically intercept the users’ uplink information. For this scenario， a maximization problem of system secrecy transmission rate based on a fairness criterion is formulated. A cooperative optimization algorithm integrating Block Coordinate Descent （BCD）， Semi-definite Relaxation （SDR）， Successive Convex Approximation （SCA） and a Dual Deep Q-network （DDQN） is also proposed. This algorithm collaboratively designs the UAV’s flight trajectory， resource allocation strategy and beamforming parameters， to balance security and communication performance. Simulation results demonstrate that the the proposed scheme offers superior flexibility and secure transmission advantages in dynamic adversarial environments， and elucidates the intrinsic correlations among the R-UAV’s onboard resources， hardware configurations and system performance.

Key words: space-air-ground networks, uplink secure transmission, dynamic adversarial environments, UAV trajectory, beamforming, resource allocation