Sensing and Recognition for Multiple-Primary-Power-Level Scenario With Noise Uncertainty

Abstract

In this paper, we consider spectrum sensing in a new cognitive radio (CR) scenario where the primary user (PU) could possibly transmit with multiple power levels and where the noise power at the secondary user (SU) side is uncertain. The target for the SU is not only to detect the presence of PU, as in the conventional “on–off” CR scenario, but to identify the PU's transmitting power level as well. With the aid of the generalized likelihood ratio test (GLRT), we prove that energy detection (ED) is still the optimal detection technique and then derive the closed-form decision thresholds and the corresponding analytical performance. Interestingly, a unique phenomenon called power ambiguity arises in this new CR scenario that demands careful investigation. Moreover, the signal-to-noise ratio (SNR) wall in the conventional binary CR also exists here and separates all power levels into two groups: One group leads to robust recognition, whereas the other does not. To make the discussion complete, we further design a cooperative spectrum sensing scheme and derive the corresponding analytical performance expression. In the end, we validate our study through various numerical results.