Attack Prevention for Collaborative Spectrum Sensing in Cognitive Radio Networks

Lingjie Duan, Alexander W. Min, Jianwei Huang, Kang G. Shin

Abstract

Collaborative spectrum sensing can significantly improve the detection performance of secondary unlicensed users (SUs). However, the performance of collaborative sensing is vulnerable to sensing data falsification attacks, where malicious SUs (attackers) submit manipulated sensing reports to mislead the fusion center's decision on spectrum occupancy. Moreover, attackers may not follow the fusion center's decision regarding their spectrum access. This paper considers a challenging attack scenario where multiple rational attackers overhear all honest SUs' sensing reports and cooperatively maximize attackers' aggregate spectrum utilization. We show that, without attack-prevention mechanisms, honest SUs are unable to transmit over the licensed spectrum, and they may further be penalized by the primary user for collisions due to attackers' aggressive transmissions. To prevent such attacks, we propose two novel attack-prevention mechanisms with direct and indirect punishments. The key idea is to identify collisions to the primary user that should not happen if all SUs follow the fusion center's decision. Unlike prior work, the proposed simple mechanisms do not require the fusion center to identify and exclude attackers. The direct punishment can effectively prevent all attackers from behaving maliciously. The indirect punishment is easier to implement and can prevent attacks when the attackers care enough about their long-term reward.

System Model and the SDF Threat

In Cognitive Radio Networks (CRNs), Secondary Users (SUs) utilize licensed spectrum when the Primary User (PU) is idle. To ensure reliability, SUs perform collaborative spectrum sensing, pooling their observations at a central Fusion Center to determine channel occupancy.

Figure 1: System model showing the interaction between the Fusion Center, honest SUs, and coordinated attackers.

This architecture is highly vulnerable to Sensing Data Falsification (SDF) attacks. The authors explore a sophisticated scenario involving cooperative rational attackers. These attackers exploit information asymmetry by overhearing honest SUs' reports before submitting their own, allowing them to coordinate lies that maximize their collective throughput. Without intervention, honest SUs are effectively crowded out of the spectrum while still incurring penalties for collisions caused by the attackers' aggressive behavior.

Proposed Prevention Mechanisms

The researchers propose two mechanisms that leverage a "smoking gun" event: a collision with the PU occurring after the Fusion Center has already declared the channel "busy." Since honest SUs strictly obey the busy signal, any collision in this state definitively proves an attack has occurred.

1. Direct Punishment

The Fusion Center imposes a significant fine ($C_b$) on all SUs whenever an attack is detected. By making the fine high enough, the expected reward for a rational attacker becomes negative, deterring the attack entirely. Interestingly, the required threshold for this fine is highest when there is only a single attacker, as they have less to lose from a collective penalty than a large group sharing the same fine.

2. Indirect Punishment

Upon detecting an attack, the Fusion Center permanently terminates collaborative sensing, forcing all SUs to sense individually. This "isolation" removes the attackers' ability to overhear honest reports, significantly increasing their risk of collisions and reducing long-term utility. This mechanism is particularly effective against attackers who value long-term rewards (high discount factor).

Key Results and Effectiveness

The effectiveness of these strategies depends on the attacker's behavior—specifically whether they are "attack-and-run" (short-term) or "stay-with-attacks" (long-term) players.

Figure 3: Summary of attack prevention outcomes for different scenarios and punishment types.

As summarized in the table above, Direct Punishment is a robust deterrent across all scenarios, completely preventing attacks by making them unprofitable. Indirect Punishment is highly effective against long-term attackers but cannot prevent "attack-and-run" scenarios where the malicious actor does not care about future collaboration.

Crucially, these mechanisms do not require the Fusion Center to identify specific malicious nodes. By shifting the focus from identification to collective accountability, the system effectively "cleans" the network by ensuring the cost of malice outweighs the benefits.

Limitations and Future Outlook

While powerful, these mechanisms assume attackers are rational (profit-driven). They may not deter "Byzantine" attackers whose sole goal is network disruption regardless of personal cost. Additionally, the model assumes the Primary User can reliably detect and communicate collisions to the Fusion Center. Future research may explore correlated PU traffic patterns and the impact of noise on the reporting channel, which could potentially trigger false punishments.

Summary updated

Abstract

Collaborative spectrum sensing can significantly improve the detection performance of secondary unlicensed users (SUs). However, the performance of collaborative sensing is vulnerable to sensing data falsification attacks, where malicious SUs (attackers) submit manipulated sensing reports to mislead the fusion center's decision on spectrum occupancy. Moreover, attackers may not follow the fusion center's decision regarding their spectrum access. This paper considers a challenging attack scenario where multiple rational attackers overhear all honest SUs' sensing reports and cooperatively maximize attackers' aggregate spectrum utilization. We show that, without attack-prevention mechanisms, honest SUs are unable to transmit over the licensed spectrum, and they may further be penalized by the primary user for collisions due to attackers' aggressive transmissions. To prevent such attacks, we propose two novel attack-prevention mechanisms with direct and indirect punishments. The key idea is to identify collisions to the primary user that should not happen if all SUs follow the fusion center's decision. Unlike prior work, the proposed simple mechanisms do not require the fusion center to identify and exclude attackers. The direct punishment can effectively prevent all attackers from behaving maliciously. The indirect punishment is easier to implement and can prevent attacks when the attackers care enough about their long-term reward.
BibTeX
@article{duan2012attack,
  author = {Lingjie Duan and Alexander W. Min and Jianwei Huang and Kang G. Shin},
  title = {{Attack Prevention for Collaborative Spectrum Sensing in Cognitive Radio Networks}},
  journal = {IEEE J. Sel. Areas Commun.},
  year = {2012},
  url = {https://openreview.net/forum?id=OpAEJ6zkUg},
}
Figure