Compression and Cooperation for Wideband Spectrum Sensing and Cognition

Michigan Technological University, Houghton MI

Investigators

Abstract

The objective of this research is to increase the resolution and reduce the cost of wideband spectrum sensing and cognition, which are challenging core issues in spectrum-sharing cognitive networks. The approach seeks to leverage the benefits of compressive sampling and user cooperation to develop a collaborative compressed sensing framework for wideband networks. Research thrusts include compressed wideband sensing at affordable signal acquisition costs and reliable collaborative sensing with low network overhead. This research combines compression and collaboration techniques with the goal of innovation in wideband spectrum sensing. The project seeks to develop a suite of compressed sensing algorithms that exploit spectral sparsity to reduce the need for high-speed sampling in ultra-wideband and wideband radios. A decentralized approach for joint cooperation and compression, which exploits spatial diversity to alleviate the hidden terminal problem caused by wireless channel fading, is investigated for ad hoc cognitive networks. Further, the research examines fundamental tradeoffs in a cooperative sensing system including diversity gain, compression, sensing time and complexity. This project has the potential to lead to techniques that can improve the spectrum utilization efficiency and user capacity of wireless networks, thus allowing a multitude of new cognitive radio transceivers. Such devices include wireless sensors and radio frequency identification (RFID) chipsets for monitoring and tracking applications. The compressed sensing and decentralized collaboration techniques being investigated also have the potential to contribute to other sensing-related applications, such as wireless sensor networks. Research experiences gained through this project help to prepare students for engineering in the 21st Century.