Spectrum sensing with active cognitive systems

Spectrum sensing is critical for cognitive systems to locate spectrum holes. In the IEEE 802.22 proposal, short quiet periods are arranged inside frames to perform a coarse intra-frame sensing as a pre-alarm for fine inter-frame sensing. However, the limited sample size of the quiet periods may not guarantee a satisfying performance and an additional burden of quiet-period synchronization is required. To improve the sensing performance, we first propose a quiet-active sensing scheme in which inactive customer-provided equipments (CPEs) will sense the channels in both the quiet and active periods. To avoid quiet-period synchronization, we further propose to utilize (optimized) active sensing, in which the quiet periods are replaced by 'quiet samples' in other domains, such as quiet sub-carriers in OFDMA systems. By doing so, we not only save the need for synchronization, but also achieve selection diversity by choosing quiet sub-carriers based on channel conditions. The proposed active sensing scheme is also promising for spectrum sharing applications where both the cognitive and primary systems can be active simultaneously.