Secure Cooperative Spectrum Sensing Using Interference Signatures

Project Description

We introduce a primary user emulation test (PUET), under which a trustful central entity (e.g., a cellular base station) transmits a test signal while other users are sensing the spectrum. The core of PUET is to correlate the reported sensing data with the transmission power of the test signal. Since this test signal is, in reality, an interference to the sensing of a primary signal, sensors cannot distinguish the test signal frma the primary signal. Considering this characteristic of sensors, PUET detects attacks by evaluating the consistency of channel parameters, which are not known to sensors. By realizing this defense mechanism, PUET checks the validity of reports from each sensor separately.

 

Objective

 

Cooperative spectrum sensing is vulnerable to sensing data falsification attacks due to the distributed nature of spectrum sensing in Cognitive Radio Networks (CRNs). As the goal of a sensing data falsification attack is to cause an incorrect decision on the presence/absence of a Primary User (PU) signal, malicious or compromised Secondary Users (SUs) may intentionally distort the measured Received Signal Strengths (RSSs) and share them with other SUs. Then, the effect of erroneous sensing results propagates to the entire CRN. This type of attacks can be easily launched since the openness of programmable software-defined radio (SDR) devices makes it easy for (malicious or compromised) SUs to access lowlayer protocol stacks, such as PHY and MAC. However, detecting such attacks is challenging due to the lack of coordination between PUs and SUs, and unpredictability in wireless channel signal propagation, thus calling for efficient mechanisms to protect CRNs.

 

Approach

In this project, we propose a mechanism, called PUET, to detect the falsification of sensing results. PUET resides at a trustful central entity (e.g., a cellular base station) that transmits a test signal while other SUs are sensing the spectrum. Since this test signal is, in reality, an interference to the sensing of a PU signal, SUs’ RSSs should be the sum power of the PU signal and the test signal. Then, the central entity can detect attacks by checking if the reported sensing data reflects the interference caused by the test signals.

 

Results

To evaluate the performance of PUET, we constructed a testbed with Universal Software Radio Peripheral 2 (USRP2) in our Department building. We deployed 5 USRP2 nodes as in the topology shown below. The testbed consists of a PU (denoted as circled P in the figure), a BS (denoted as circled B in the figure), and 3 CPEs (denoted as circled C in the figure).

The BS determines the presence/absence of PU signals by collecting sensing reports that are not filtered out by PUET. As a baseline data-fusion scheme, we use the majority rule for this experiment. Below shows the performance of detecting PU signals with and without PUET: “MAJ” represents the majority-rule based data-fusion, and “PUET” represents application of the majority rule after PUET filtered out attacks separately.

 

Participants

Faculty

Student

  • Seunghyun Choi

 

Publications

  • Seunghyun Choi and Kang G. Shin. Secure Cooperative Spectrum Sensing in Cognitive Radio Networks Using Interference Signaturesaccepted for IEEE Conference on Communications and Network Security, 2013.