Spectrum opportunity forecasting for energy efficient sensing in cognitive radio networks

Energy efficiency in cognitive radio networks (CRN) is of paramount importance since secondary users (SU) are often likely to be energy constrained. While spectrum sensing (SS) is a critical CRN function, repetitive SS events can significantly reduce the battery life of sensing devices. However, energy efficiency can be improved by employing spectrum opportunity forecasting (SOF) and optimal scheduling for sensor node activation to reduce the required number of SS events. In this paper, a combination of SOF and optimal scheduling is thus explored. The application of SOF to cooperative SS is one of the unique contributions of this paper. Results indicated that this combination could significantly reduce sensor node energy consumption and in so doing increase CRN lifetime. A positive linear relationship was found between forecast length and CRN lifetime. A very large improvement in the collective lifetime of the sensor network was observed when SOF was employed (roughly 21.2 time samples longer than when only optimal scheduling was employed) for a forecast length of ten SS events, using implicit enumeration under the uniform sensor node distribution.