Performance evaluation of directional spectrum sensing using an over-the-air testbed

Cognitive radio is an emerging approach to improve the spectral utilisation, and can also be applied in self-organized mobile networks. Spectrum sensing to learn from the radio environment plays a key role of intelligent radios.

To further enhance the self-organized capabilities of cognitive radio nodes, we propose to exploit the spatial resource as a further degree of freedom using antennas with orthogonal radiation patterns. We built a testbed based on a software-defined radio and an over-the-air test environment, to analyse the performance of an energy-based sensing algorithm in multiple bands and different directions. This technique enables reconstruction of primary and secondary user activities in an authentic radio environment.

Initial measurements with a multi-band and directional antenna array at the device side were carried out, using eight wideband antennas in a circular arrangement at the over-the-air emulation side. These antennas were fed with static wideband signals at different carrier frequencies, emulating a line-of-sight scenario of eight primary users. The results show that the energy-based sensing algorithm is able to draw valid conclusions on user activity in different bands in different directions, indicating a possible multiple use of frequencies.