Bandwidth-Adaptive Waveforms for Dynamic Spectrum Access Networks

This paper presents a novel solution to the challenge of waveform bandwidth estimation in DySPAN systems using intentionally embedded cyclostationary signatures. Dynamic spectrum access networks (DySPANs) achieve high levels of spectrum use efficiency by exploiting spectrum white spaces, radio frequencies which are unused at given times and places. These white spaces have bandwidths which may not be known in advance and so, in order to fully exploit them, a DySPAN system must employ bandwidth-adaptive waveforms. It has been shown that multi-carrier transmission schemes can be leveraged to generate waveforms of flexible bandwidth. However, robust bandwidth estimation techniques are required for receiving nodes to successfully detect these signals and establish communication links. This paper makes three key contributions. Firstly, a novel technique for the generation of bandwidth-adaptive multi- carrier waveforms is proposed. Secondly, two approaches for parametrization of these bandwidth-adaptive waveforms in receiving DySPAN nodes are presented. The advantages and disadvantages of both bandwidth estimation techniques are examined and comparisons are drawn with approaches adopted for the IEEE 802.16 and draft IEEE 802.22 standards. Finally, the performance of both techniques under multipath fading conditions are examined using simulation results.