Adversary analysis of cockroach network under Rayleigh fading channel: Probability of error and adversary detection
Wong, Tze C.
AdvisorKwon, Hyuck M.
MetadataShow full item record
This paper extends the design of a cockroach network from a wire network, without considering thermal noise and channel fading, into a wireless Rayleigh fading channel. This research is developed and split into two directions: probability of error for data transmission and probability of detection for adversary node. A lookup table is proposed in order to speed up the process of data decoding and also to identify the node that has the highest probability of behaving as an adversary. This table is a summary of all combination data received at the destination so that a decision can be made logically. In the section relative to probability of error, the analysis begins with deriving the equation of probability of error between the source, nodes, and destination. Then, by taking the approximation, the probability of each data combination at the destination is determined. With the aid of the lookup table, the probability of error can be obtained by summing the combination causing the error. Similar to the probability of error, the probability of detection is also determined with the assistance of the lookup table by using a combination of signals received at the destination. By summing up the probability of this combination, the probability of detection and false alarms can be obtained, with and without the existence of an adversary. In the end, the simulation result is compared to the derived equation of probability of error and detection. Both analysis and simulation results show that the probability of error achieves 10-2, when the signal-to-noise ratio (SNR) is about 15 dB under the condition of no adversary and when the SNR is in the range of 19 to 24 dB when one of the nodes is compromised. On the other hand, the probability of false alarm is reduced significantly when the SNR is higher than 20 dB, and the rate of successful adversary detection is about 95% at 20 dB.
Thesis (Ph.D.)-- Wichita State University, College of Engineering, Dept. of Electrical Engineering and Computer Science