Proactive Eavesdropping With Jamming Power Allocation in Training-Based Suspicious Communications

This letter studies proactive eavesdropping with one legitimate monitor (E) in the classical single-hop suspicious communication. Specifically, unlike all previous works that ignored the suspicious channel training phase and just considered the jamming power optimization of E in the suspicious data transmission phase to facilitate eavesdropping, this letter advances the research by comprehensively investigating the jamming power allocation of E in both phases, with the purpose of maximizing its eavesdropping success probability. Under this setup, we first derive an exact expression of the objective and reveal the existence of a fundamental trade-off in deciding the jamming power allocation, for which a simple one-dimensional search is employed to find the optimal solution. To simplify the analysis, we further derive a tight approximation of the objective and then develop a very fast alternating optimization algorithm to find the sub-optimal jamming power allocation. Moreover, we extend our analysis to the case where the channel state information is available at the suspicious transmitter via channel feedback. Simulation results demonstrate the effectiveness of our proposed scheme compared to competitive benchmarks.