Location based spectrum sensing in cognitive radio systems

Abstract

Cognitive radio is a promising solution to the spectrum utilization problem in time and space. One of the most critical points of cognitive radio is spectrum sensing. A reliable sensing is important for not disturbing the primary user and utilizing the spectrum usage. In this thesis energy detection algorithm is selected because of its ease of implementation, low cost and low computational complexity. Secondary user knows the location information of the primary users. A central database which is called Radio Environment Map (REM) stores the location information of the primary users, and on which frequencies each primary user operates. The REM calculates the distances between the primary users and the secondary user. Using the distance information and the inverse proportion relation between distance and signal to noise ratio (SNR), the secondary user decides on which sequence it will sense the primary users in order to maximize its signal to noise ratio. The secondary user senses the primary users beginning from the closest to the furthest, until it finds an empty channel. It stops sensing when it finds a non transmitting primary user. Using this sequence it will maximize average SNR and increase sensing performance. The error capacity of this method is also analyzed. Performance of proposed method under erroneous location information is measured with simulations. Another method is proposed when the maximum interference distance of secondary user is known. By this new information the CU can use the channels of the PUs which are outside the interference range without sensing them. The final proposed algorithm is cooperation of two cognitive users using the distance information of primary users. Each cognitive user starts sensing from the closest PU to itself, and stops sensing when one of them finds an empty channel or the next closest PU that is closer to the other CU.