Security Enhanced Clustering Scheme for Routing in Cognitive Radio Sensor Networks

Due to the scarcity of wireless spectrum, it entails the utilization of dynamic shared spectrum solutions, which can be availed with an intelligent wireless communication technology called Cognitive Radio (CR). Wireless sensor nodes equipped with CR reconfigure dynamically to utilize the vacant spectrum holes, forming a Cognitive Radio Sensor Network (CRSN), which has high demand in Internet of Things (IoT) applications. Despite of attaining high spectrum utilization via CR capabilities, CR based sensor networks require energy efficient solutions due to their resource-constrained behavior. Therefore, a clustered architecture becomes suitable for maintaining opportunistically available radio resources and energy efficiency in CRSNs. In this paper, a trust-based secure clustering scheme is introduced for routing in CRSNs, where both direct and indirect/recommended trust are considered, along with an aging method for updating trust. This scheme accumulates the history of trust values to detect and eliminate malicious nodes from the network. Additionally, during cluster formation, two parameters weight value and neighbor set of each node are taken into consideration. These parameters are obtained through the suggested neighbor selection scheme. The neighbor selection approach considers the neighbor’s requirements in terms of transmission time and the availability of a common channel between the Cognitive Sensor (CS) node and the associated neighbor set. Thereafter, the proposed secure clustering approach will incorporate cluster head rotation to ensure an even distribution of energy consumption among the nodes in a cluster. Simulation based study demonstrates that the proposed scheme enhances detection efficiency, detection ratio, the network’s relative average packet delivery ratio, relative average packet drop ratio, average end-to-end delay, and average energy consumption of the network for selective forwarding, blackhole, and ON-OFF attacks in comparison with an existing method. Moreover, the correctness of the proposed scheme is evaluated in terms of precision and F1-Score.