Relay-Aided Proactive Eavesdropping with Learning-Based Power and Location Optimization.

This work investigates a relay-aided proactive eavesdropping network, where a legitimate monitor eavesdrops on communication between suspicious users via a friendly amplify-and-forward full-duplex (FD) relay. The closed-form expressions are derived, including the decoding outage probability for the suspicious link, the eavesdropping non-outage probability for the eavesdropping link, and the average eavesdropping rate (AER). To maximize the AER, separate optimization problems for the power and location of the relay are settled, employing the good fitting characteristic of the deep feedforward neural network. Then, a low-complexity learning-based iteration algorithm is proposed to solve the joint optimization problem. Numerical results demonstrate the effectiveness and optimality of the proposed algorithm, and show that the optimized FD relay-aided proactive eavesdropping scheme outperforms the existing benchmark schemes in terms of the AER.