The goal of this project is to devise an efficient framework for achieving radio frequency (RF) spectrum awareness through generating reliable and dynamic radio environment maps (REMs). This is a notable step towards the proliferation of realistic cognitive radio networks (CRNs), which enable spectrum sharing to alleviate the problem of spectrum scarcity. This efficient use of spectrum empowers ever-increasing applications and services with a great impact on national health, welfare, public safety, and economic growth. In addition, the developed data analysis tools provide a distinctive solution to high-dimensional signal sampling and processing, enabling further development in a wide range of applications, such as big data, Internet of Things, and wireless sensor networks, which can promote social and economic progress. This project integrates research and education through new course development and revisions, and involving underrepresented minorities, graduate, and undergraduate students in research.
This research bridges the gap between the theoretical research in tensor data analysis and wireless communications and networking to facilitate spectrum awareness. The investigators will develop a tensor-based framework to generate dynamic and reliable REMs that include RF signal power distribution over space, time and frequency. Tensor-based analysis facilitates the integration of the inherent properties and data structures that exist in CRNs as well as the prior knowledge of the behavior of the primary network. Bayesian and structure-based tensor decomposition is investigated, for which existence and uniqueness conditions are studied. The outcome of the decomposition is employed to obtain a model-based interpolation of sensor readings and to generate the REM. Corresponding to an REM, a reliability map will be created and utilized for optimal joint spectrum sensor and channel selection.