Due to the strict management policy and limited space in wireless spectrum, it is very difficult to overcome the demands of high data rate and bandwidth requirements in the wireless communication. To deal with this problem effectively, random allocation of the spectrum is considered, which resulted in the concept of cognitive radio network (CRN). Resource Allocation and Spectrum sensing in CRN is of high interest, where opportunistic users also called secondary users (SUs), have to detect the licensed primary user (PU) spectrum and make use of the vacant. The effects of multipath fading, shadowing and receiver uncertainty lead to poor spectrum sensing performance of individual users. Cooperative spectrum sensing (CSS) is a solution to acquire accurate information about the PU channel in the fading and shadowing environment. CSS enables each user to share its local sensing information with the neighbors to reach a more precise spectrum sensing decision. The malicious users (MUs) false sensing reports prevent the fusion center (FC) from taking a precise final decision, hence it can reduce effectiveness of CSS system. Many detection and suppression schemes are found in the literature to make the FC decision secure and robust in the presence of these abnormalities. This dissertation is a contribution to the above mentioned areas. The dissertation is mainly divided into three parts. In the first part, we have proposed two variants of the Kullback Leibler (KL) divergence, including simple KL divergence and weighted KL divergence schemes to prevent the system from always yes, always no, opposite and random opposite categories of MUs without identification. The final decision made by the FC, using simple KL divergence and weighted KL divergences schemes is more precise with high detection, less false alarm and low energy consumption. In the second part, we have proposed heuristic algorithms, including Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) based soft and hard fusion combination schemes at the FC. In the last part, for efficient detection and mitigation of MUs, we have proposed statistical techniques. In this section, FC is allowed to take its cooperative decision normally about the sensing channel, based on the received local decisions of the cooperative SUs. When enough statistics are collected about the reporting users, Box-whisker‟s plot (BWP) and Hampel‟s test (HT) are employed to detect and separate the false sensing data provided by MUs as abnormal data and is able to further shape the hard and soft fusion decisions based on the reported data of the normally reporting users. The effectiveness and reliability of our proposed techniques are demonstrated in the results and simulations where graphs are plotted for the detection, false alarm, miss-detection and error probabilities against different types of MUs, total number of cooperative users and signal to noise ratios (SNRs). The spectrum sensing responsibility in the presence of various categories of MUs is a challenging job that is made authentic using KL divergence, GA, PSO and some statistical techniques in the dissertation. The proposed techniques in the dissertation allow the FC to estimate the PU channel status accurately so that the SUs are able to make use of the available spectral holes without any disturbances and interference to the legitimate users. In the industrial environment, sensors and robot in coordination detect the abnormal behavior of any robot, as the malfunctioning in such robots due to any reason reduce overall performance of the system. Therefore, the proposed CSS model can precisely detect faulty sensors and robots in the industrial environment and it has a centralized performance monitoring mechanism.