Cognitive Radio (CR) technology has been proposed in order to make efficient usage of spectrum. In this technology, users who have no spectrum licenses, also known as secondary users (SUs), are allowed to use temporarily unused licensed spectrum. CR is based on effective spectrum sensing. Some environmental consequences such as multi-path fading, shadowing and hidden terminals affect the results of spectrum sensing obtained by CR users. Cooperative Spectrum Sensing (CSS) is suggested to decrease such consequences. In CSS, CR users share their information obtained during individual spectrum sensing and make cooperative decisions which is more accurate than individual decisions. The cooperation among CR users raises new concerns on the reliability and the security in cooperative sensing. This is because, when multiple CR users cooperate in sensing, it is possible some CR users which are malicious users (MUs) report unreliable or falsified sensing data which can easily influence the cooperative decision. Among the methods presented to detect and eliminate the effect of MUs, outlier based method can be pointed. The sensible definition of an outlier is an observation that highly deviates from other observations makes suspicion that such an observation has been generated by a different mechanism. In this thesis, we the Local Outlier Factor (LOF) and Simple Local Outlier Factor (SLOF) have been used for eliminating the MUs Data. In this thesis to defense against MUs cooperation schemes have been presented. In such schemes, energy detection and cyclostationary feature detection have been implemented. Then in order to remove the MUs data, LOF and SLOF values are allocated to each of sensed data and those data whose LOF and SLOF are greater than a threshold value, are identified as MUs data and removed from combination process. It is shown that identifying MUs based on the proposed schemes does not depend on detection method. Furthermore, the proposed schemes do not require any other pre-knowledge about data distribution, primary network, location of primary transmitter and location of secondary users. In other words, the proposed schemes do not require to keep the history of CR users behaviors and are able to detect and eliminate several MU. Keywords : Cognitive radio Networks; cooperative spectrum sensing; malicious user detection in cooperative spectrum sensing