Abstract :
Cognitive radios must examine the contents of the RF bands potentially available for communication prior to commencing transmission. This process is often referred to as spectrum sensing, and its goal is the declaration of subbands as either occupied (black space) or unoccupied (white space). If the radio is equipped with an OFDM waveform, even a set of irregularly spaced narrow white subbands can be useful by properly selecting the subset of subcarriers on which to transmit. It is important, therefore, to provide the radio with the ability to automatically detect all available white space for a wide variety of RF scenes: densely packed emitters, arbitrarily colored background noise, mixtures of emitter bandwidths, etc. In other contexts, such as with the wideband single-carrier MUOS (Mobile User Objective System) satellite communication system, the black spaces are avoided by an adaptive multi-notch transmit filter whose passbands reflect the detected white spaces. In this paper, a novel automatic multi-resolution general-purpose white-space detector is described. To illustrate the difficulty of the problem and the potential of the algorithmic solution, the method is applied to simulated and collected UHF data sets.
