Detection of radio receivers: an experimental evaluation approach

Abstract

This paper considers the problem of detecting extremely low power emissions due to local oscillator (LO) leakage. This challenging problem arises in cognitive radio (CR) networks when trying to detect a primary user in reception mode (the hidden node problem), and also in physical layer security applications when trying to determine the presence of an eavesdropper. In contrast to the numerous works on the detection of an active radio transmitter, the problem addressed in this paper still requires further research. As a matter of fact, previous studies focus on the theoretical analysis of the detection strategies, and the reported performances do not take into account the additional limitations imposed by the existing hardware technologies. In this paper, we provide an experimental evaluation of two different detectors based on the averaged periodogram or the total energy, respectively. Moreover, we propose practical guidelines to overcome the hardware limitations and maximize the detection performance. Our experimental measurements using Universal Software Radio Peripheral (USRP) boards reveal that the widely employed energy detector is outperformed by the periodogrambased detector.