Wireless Secrecy in Cellular Systems with Infrastructure--Aided Cooperation

In cellular systems, confidentiality of uplink transmission with respect to eavesdropping terminals can be ensured by creating intentional inteference via scheduling of concurrent downlink transmissions. In this paper, this basic idea is explored from an information-theoretic standpoint by focusing on a two-cell scenario where the involved base stations are connected via a finite-capacity backbone link. A number of transmission strategies are considered that aim at improving uplink confidentiality under constraints on the downlink rate that acts as an interfering signal. The strategies differ mainly in the way the backbone link is exploited by the cooperating downlink- to the uplink-operated base stations. Achievable rates are derived for both the Gaussian (unfaded) and the fading cases, under different assumptions on the channel state information available at different nodes. Numerical results are also provided to corroborate the analysis. Overall, the analysis reveals that a combination of scheduling and base station cooperation is a promising means to improve transmission confidentiality in cellular systems.

💡 Research Summary

This paper investigates a novel physical‑layer security technique for cellular networks that exploits simultaneous uplink and downlink transmissions together with backhaul‑based cooperation between neighboring base stations (BSs). The basic premise is that, while a mobile terminal A sends a confidential uplink message to its serving BS B, the adjacent BS C concurrently transmits a downlink signal to its own user D. Because the downlink transmission reaches the eavesdropper E, it acts as intentional interference (artificial noise) that degrades E’s ability to decode A’s message. The key to preserving the legitimate link quality is to let BS C convey some information about its downlink waveform to BS B over a finite‑capacity wired (or optical) backbone link of capacity (C_L). With this side information, B can partially or fully cancel the interference caused by C, thereby improving the secrecy rate.

The authors first treat the large‑backbone case ((C_L \ge R_C), where (R_C) is the downlink rate). In this regime C can forward the entire downlink codeword to B, enabling perfect interference cancellation at B. Consequently the legitimate link capacity remains (C(\gamma_{AB})) while the eavesdropper sees a degraded channel whose effective SNR is reduced by the presence of C’s signal. The achievable secrecy rate is \