Cognitive Radio Networks (CRNs) were proposed to solve the problem of spectrum scarcity, low utilization of spectrum resources and increasing demands for wireless communication services. In these networks, users are divided to two categories called Secondary Users (SUs) and Primary Users (PUs). The frequency spectrum belongs to PUs and SUs are allowed to access the spectrum resources of licensed PUs without causing harmful interference with them. Cognitive Radio Sensor Networks (CR) as a subset of CRNs are a type of Wireless Sensor Networks (W) where sensors are implemented with cognitive radio capability. The sensors in such networks can use the frequency spectrum of PUs as the SUs in CR. CR with interweave structure are the networks where sensors vacate the frequency spectrum when the presence of PUs is detected on the frequency spectrum. The W with industrial automation application and target tracking are the examples of such networks. To make sensors aware of the presence of PUs has a special importance in such networks. Sensors should perform spectrum sensing to become aware from the presence of PUs. In order to improve the reliability of decisions made by sensors about the presence of PU, Cooperative Spectrum Sensing (CSS) is used. In CSS, several sensors independently decide about the presence of PU(s) and the final decision about the presence of PU on the frequency spectrum is made based on their decisions. One of the important challenge of CSS is energy consumption for sensors. In this thesis, three system models are considered for CR with respect to the location of PU(s) and sensors. In the first model, the frequency spectrum belongs to one PU which is located far from sensors. In such a model, all sensors receive the same SNRs from the PU. In the second model, the frequency spectrum belongs to one PU which is located close to sensors such that sensors receive different SNRs from the PU. In the third model, the frequency spectrum belongs to more than one PU. The received SNR for each sensor from a PU is different from its received SNR from other PU. In all of three models, the reporting channels between Fusion Center (FC) and sensors are considered with different error probabilities and sensors have various distances from FC. Then, new algorithms are proposed for these models to engage sensors in CSS process with respect to their energy constraints and provide the desired sensing accuracy. These algorithms have been compared with the most recent research works in the field of energy consumption reduction for CSS and the analysis and simulation results confirm that the new algorithms are able to reduce energy consumption and increase the network lifetime. Keywords : Energy consumption, Sensor selection, Spectrum sensing, Cognitive radio.