Jamming, broadcasting to intentionally interfere with wireless reception, has long been a problem for wireless systems. Recent research demonstrates numerous advances in jamming techniques that increase attack efﬁciency or reduce the probability an attack will be detected by choosing attack parameters based on a system’s conﬁguration. In this work, we extend the attacker’s capabilities by modifying the attack parameters in response to the observed performance of the target system, effectively creating a feedback loop in our attack model. This framework allows for more intricate attack models that are tuned online allowing for closer to optimal attacks against legitimate systems. To show the feasibility of the listening and attacking framework we introduce an attack called Self-Tuned, Inference-based, Real-time jamming or STIR-jamming. This attack listens to legitimate communication trafﬁc, infers the systems performance, and optimizes jamming parameters. We propose the two types of STIR-jamming, mSTIR-jamming and tSTIR-jamming, and implement these attacks against an IEEE 802.15.4 link as a case study. With the empirical results, we demonstrate the attack system adapting to various scenarios and ﬁnding stable solutions.
Bruce DeBruhl, Yu Seung Kim, Zachary Weinberg and Patrick Tague. "STIR-ing the Wireless Medium with Self-Tuned, Inference-Based, Real-Time Jamming" 9th IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS)
Available at: http://works.bepress.com/patrick_tague/10/