DETECTION AND ISOLATION OF PRIMARY USER EMULATOR IN COGNITIVE RADIO NETWORK USING HYBRID OF ANGLE OF ARRIVAL AND RECEIVED SIGNAL STRENGTH

Abstract

Cognitive Radio (CR) technology is the candidate panacea to the problem of spectrum scarcity in the wireless world. However, this emerging technology is faced with security challenges. The most severe among these security challenges is Primary User Emulation Attack (PUEA). One of the methods to detect Primary User Emulator (PUE) is via localisation, of which there are two major categories: range-free and range-based. Range-free localisation is cost effective, less computationally complex and easy to deploy. However, it is less accurate when compared with range-based category. Since accuracy is fundamental in localisation, range-based localisation scheme was adopted in this work. The range-based category is reported to be more accurate although with higher complexity. Among this category are Angle of Arrival (AOA), which utilises angular measurements to localise the PUE, and the Received Signal Strength (RSS), which employs only distance to localise the PUE. To improve performance of range based methods, this research hybridised AOA and RSS techniques to localise PUEs in television (TV) white spaces. This scheme determines the angle at which the Primary User’s (PU’s) signal arrives at the Secondary Users (SUs) and the distance between the PU and SUs in the Cognitive Radio Network (CRN). Because in a TV white space, the location of PU is known, the computed AOA and the distance obtained from the RSS are therefore used to determine the position of a PU’s signal transmitter. This position is compared with the location of the PU to ascertain the true source of the signal, thus detecting the PUE. Computer simulations demonstrated that the hybrid scheme estimated the position of the PUE much faster and with a much lower Root Mean Square Error (RMSE) of 5.00x10-3 after 20 iterations. This greatly outperformed RSS and AOA methods that estimated the position of PUE after 50 iterations with RMSE of 2.00x10-1 and 1.00x10-2 respectively when considered individually. Furthermore, investigation was made on the selection of the best pair of SUs to be used in the detection processes. It was discovered that a pair of SUs from the same communication environment whose RSS values are very close, detected PUE better (with RMSE of 4.7x10-3 after 20 iterations) than a pair of SUs whose RSS values are higher but in different communication environments as they localised PUE with RMSE of 6.0x10-3 after 70 iterations. The significance of this result is appreciated especially when attention is given to the fact that speed, accuracy and energy efficiency are essential in the efficient operation of cognitive radios. Energy-efficient operations are essential in the current global energy crises that wireless systems face. Moreover, by isolating the detected PUE from Cognitive Radio Network (CRN), there is availability of more spectrum holes that will accommodate newer wireless technologies for effective communication. Furthermore, Secondary Users (SUs) have more transmission time, improved quality of service (QoS), connection reliability, higher throughput and improvement in the overall general performance of the entire cognitive radio network.