Abstract

Connectivity in remote areas continues to be a major challenge despite of the evolution of cellular technology. 5th Generation (5G) technology can address remote connectivity if lower carrier frequencies are available, which calls for shared use of spectrum to enable cost-efficient license-free solution. Therefore, spectrum sensing has its own role in future wireless systems such as mobile 5G networks and Internet of Things (IoT) to complement database approach in dynamic spectrum utilization. In this paper, a windowing based (WIBA) blind spectrum sensing method is studied. Its performance is compared to the localization algorithm based on double-thresholding (LAD) detection method. Both the methods are based on energy detection and can be used in any frequency range as well as for detecting all kind of relatively narrowband signals. Probability of detection, relative mean square error for the bandwidth estimation, and the number of detected signals were evaluated, including multipath and multi-signal scenarios. The simulation results show that the WIBA method is very suitable for future 5G applications especially for remote area connectivity, due to its good detection performance in low signal-to-noise ratio (SNR) areas with low complexity and reasonable costs. The simulation results also show importance of the used detection window selection since too wide detection window degrades the detection performance of the WIBA method.