Cognitive radio networks (CRNs) are envisaged to alleviate the spectrum and capacity shortage aggravated by the static spectrum allocation. They are crucial for meeting the requirements of wireless traffic explosion emerging with new wireless applications and services. However, the security of CRNs is critical for their practical usage and widespread proliferation. Therefore, the fundamental mechanism of spectrum sensing in these systems has to be investigated from the security perspective. In this work, we define and analyze coordinated attack models for Cooperative Spectrum Sensing (CSS) in CRNs. We devise a simple yet effective countermeasure scheme based on trust management for cooperative sensing. We also provide experimental results to illustrate the performance of these systems under coordinated attacks.