A Combined Spectrum Prediction and Sensing Approach for Cognitive Radios

Cognitive Radio (CR) is a promising technology to overcome the scarcity of radio frequency spectrum towards the increasing demands in wireless communication. Spectrum sensing helps us to observe the status of spectrum occupancy. To reduce the processing delays caused by spectrum sensing and to improve the efficiency of spectrum utilization, spectrum prediction algorithms were introduced. In this paper, we propose Holt’s exponential weighted moving average approach to improve the accuracy of spectrum prediction. The performance of the proposed approach is compared with existing spectrum prediction methods named Bayesian inference based spectrum predictor and exponential weighted moving average based spectrum predictor in different scenarios. We used an energy detector for spectrum sensing to obtain the observations for prediction. Here we also discuss a combined spectrum prediction and sensing approach to reduce the interference caused by the secondary user.