Cognitive Radio Sensing Architecture and a Sensor Selection Case Study

The function of a cognitive radio system (CRS) can be classified as cognitive engine (CE) and sensing. The CE fulfills spectrum analysis and decision making of spectrum access. The sensing function is realized by either a CR node or distributed spectrum sensors. In this paper, we propose a sensing architecture, which adds new functional entities, sensing controller (SC) and sensing information processor (SIP) inside CE and sensing, respectively. Based on the architecture, we give a use case where sensing is realized by distributed sensors. Selection of these sensors is performed in Rayleigh fading channel to improve the detection performance and save power consumption. Using analytical analysis we prove the benefits of sensor selection. Results show that using energy detection with 5000 symbols in - 15 dB signal-to-noise ratio (SNR) environment, we can reduce the false alarm rate from 0.95 to 0.1 while keeping 99% protection to primary users by selecting among ten sensors. Results also show that it is beneficial perform selection based on user channel.