Ubiquity of wireless connectivity and its constant growth in application and scale necessitate extremely efficient usage of spectrum as the main commodity in wireless industry. Highly regulated spectrum access policies and growing demands for spectrum access have promoted the notion of accommodating the needs of unlicensed services when the license-holding ones under-utilize the spectrum. The goal of this proposed project is to characterize the fundamental limits of spectrum sensing while recognizing the system-level and physical-level limitations imposed by costs of information acquisition, data traffic demands, and model uncertainties with sound guarantees on agility, responsiveness, and adaptability to the geo-temporal spectral occupancy variations.
Intellectual merit:
The proposed research lies at the crossroad of wireless communication, sequential statistics, and stochastic optimization. It aims to find the fundamental limits of spectrum sensing and delineates the interplay among the performance guarantees on the decisions about spectral activities, cost of sensing, communication delay tolerance, and the uncertainty level about model dynamics. This leads to novel spectrum sensing procedures in which an optimal balance among can be struck among different aspects based on the operational needs.
Broader impact:
The research has multiple technological and societal impacts. It furnishes a baseline for assessing the competency of different available spectrum sensing procedures within the same framework, and also provides novel sensing procedures that yield performance levels that match the theoretical fundamental limits. The theoretical models and concepts that are developed in this project have the potential to impact adjacent research domains, especially those involving experimental design and hypothesis testing for high-dimensional data.
