Compressed Localization and Spectrum Sensing for Cognitive Radio and Distributed Radio Surveillance

点击翻译按钮获取中文摘要

The tasks of detection, identification, and localization of radio emissions in unknown or partially unknown radio environments are crucial in a number of applications. In this project, we consider a highly agile system for distributed radio surveillance that can be deployed as a sensing system in CR networks or a system for passive radio monitoring and radio surveillance for law enforcement, public safety and regulatory purposes. The considered system is comprised of spatially distributed sensing nodes equipped with different combinations (one or more) of the sensing capabilities, e.g., wideband spectrum sensing, DoA estimation, GPS synchronization for TDoA estimation, etc., and a central unit for control and data processing connected to the nodes via communication data links.Traditional approaches for spectrum sensing, emitter localization and signal intelligence are based on the conventional Nyquist rate sampling and processing in frequency, time, and space. This often translates into stringent requirements on the sensor hardware and processing capabilities as well as required throughput of communication links. Recently emerged paradigm of compressed sensing (CS) provides a mathematical framework for simultaneous sensing and compression of signals that have sparse or compressible representation that is often the case in the considered system. The application of CS at various stages of data collection and processing allows relaxing the hardware requirements at the SNs and reducing the amount of exchanged information. CS framework however still has a significant lack of knowledge on the sub-Nyquist acquisition methods for analog signals and signal processing from received compressed samples that take into account practical constraints including robustness to noise. The main goal of this project is to bridge this gap and investigate CS-based approaches for distributed radio surveillance under realistic signal and propagation conditions.In the first phase of the project, we have applied sub-Nyquist sampling at the sensor level according to the considered sensor types. Additionally, we have investigated principle CS-based signal processing operations within each of the system tasks. In the second phase, we consider approaches for an integration of several sensing capabilities in one sensor and sensor data fusion for compressive joint time/frequency-spatial signal localization. Furthermore, we extend Software Defined Radio-based proof-of-concept demonstrations developed in the first phase according to the new findings.This project is a continuation of the CLASS project embedded within the Framework of the German-Colombian Collaborative Research Initiative in Electrical Engineering (GeCoCo-EE), which is based on the Memorandum of Understanding (MoU) between Deutsche Forschungsgemeinschaft e.V. (DFG), Germany and Departamento Administrativo de Ciencia, Tecnologia e Innovación (COLCIENCIAS), Colombia.