Exact Performance of Cooperative Spectrum Sensing for Cognitive Radios with Quantized Information under Imperfect Reporting Channels

Abstract

Spectrum sensing is the first and main step for cognitive radio systems to achieve an efficient use of the spectrum. Cooperation among cognitive radio users is a technique employed to improve the sensing performance by exploiting the diversity between the sensing channels to overcome the fading and shadowing effects which allows reduction of miss-detection and false alarm probabilities. information can be exchanged between cooperating users in different formats from the binary hard information to the full soft information. Quantized information has shown its efficiency as a trade-off between binary hard and full soft for other cooperative schemes, in this paper, we investigate the use of quantized information between cooperating cognitive users. We derive closed-form expressions of the cooperative average false alarm and detection probabilities over fading channels for a generalized system model with not necessarily identical average sensing Signal-to-Noise Ratio (SNR) and imperfect reporting channels. Numerical simulations allow us to conclude a trade-off between the quantization size and the reporting energy in order to achieve the optimal cooperative error probability.