Weighted Adaptive Spectrum Sensing for Cognitive Vehicular Networks

Abstract

Vehicular Network is an advanced application which enables smarter management of traffic and ensures safer, and more coordinated travel. The US Federal Communication Commission (FFC) has designated 75MHz of spectrum for vehicular communications (V2V and V2I). As the number of vehicles on the road increases, the spectrum available for vehicular networks becomes congested. Spectrum scarcity is a major problem in vehicular networks which needs to be addressed for improved performance. Cognitive Radio (CR) technology is a promising solution which is used to solve the scarcity of the spectrum through Dynamic Spectrum Access (DSA). CR technology integrated with Vehicular Networks (CVN) aids in the complete exploitation of underutilized licensed spectrum. Sensing the state of the channel is a fundamental step in CR networks which is carried out by the CR/Vehicular Users (VUs). However, sensing of spectrum is a challenging issue due to the continuous movement of vehicle (mobility), which results in a dynamic network topology. These challenges make the decision made by an individual VU uncertain. In the proposed work, we introduced a weight based adaptive cooperative sensing approach to enhance channel detection probability in Cognitive Vehicular Networks. The weights are calculated by considering certain parameters which includes probability of VU being outside the protection range of the PU, probability of PU being within the sensing range of the VU, Signal to Noise Ratio (SNR), velocity of the VUs, distance between VU and PU and the time of sensing. The calculated weight for the decision made by VU indicates the reliability of the signal received. A global decision is made by combining the weighted individual decisions received from VUs by applying the majority fusion rule. In addition to this, a reward-based channel recommendation for sensing is done to make the system more viable for wideband sensing. The rewards to the channels are given based on the previous states of the channels. The computational results suggest that the proposed technique outperforms conventional spectrum sensing approaches.