Visible to the public Constrained Adaptive Jamming and Anti-Jamming


Jamming has long been a problem in wireless communications. However, jamming attack models are still relatively primative, and defensive techniques can only be as strong as attack models. We therefore study the problem of enhancing jamming attack and defense models with respect to modern radio capabilities. On the offensive side, we propose a new jamming attack capability using observations about the target system as feedback to continually adapt jamming signal parameters, allowing an attacker to finely tune the attack to match a desired goal without relying on a precise model of the system. We refer to our technique as STIR-jamming, for Self- Tuned Inference-based Real-time jamming. On the defensive side, we demonstrate that certain classes of efficient and stealthy jamming attacks based on very narrow-band signal generation can be detected and mitigated using digital signal processing. We design a modified reciever architecture in which a component of the radio initiates a filter iteration process when a certain performance statistics falls below a pre-defined threshold. The iterative process continues until either the performance recovers or no reasonable solution is found, in which case the radio reports the detection result to a higher-layer service without a mitigation solution. Both our offensive and defensive techniques are currently implemented in software-defined radio, and a demo (live or video) can be made available upon request.

Award ID: 1149582

Creative Commons 2.5

Other available formats:

Constrained Adaptive Jamming and Anti-Jamming
Switch to experimental viewer