DRL-Based Joint Relay Selection and Power Allocation Strategy in Full-Duplex Eavesdropping Channels

Full-duplex technology enables the simultaneous transmission of information on the same frequency, thereby theoretically enhancing spectrum efficiency and catering to the requirements of future secure communications. However, the inherent openness and broadcast nature of wireless communication systems significantly compromise the security of data transmission in the presence of eavesdroppers. In order to optimise the safety capacity of the channel and strengthen the safety and reliability of the full-duplex system, we propose a novel strategy that integrates relay selection and power allocation using deep reinforcement learning (DRL). In situations where multiple relay nodes are involved, DRL-based selection methodology is adopted to identify the most suitable relay node. Subsequently, under a joint power constraint, the power allocation scheme ensures optimal distribution of power between the source node and the selected relay node in pursuit of achieving the specified objectives.