SpecEES: Spatio-Spectral Sensing with Wideband Feature Extraction Arrays

Cornell University, Ithaca NY

Investigators

Abstract

SPASS: Spatio-Spectral Sensing with Wideband Feature Extraction Arrays The ever-growing demand for higher data-rates combined with the rapid increase in the number of wireless devices driven by the Internet of Things (IoT), will inevitably cause congestions in the Radio-Frequency (RF) spectrum. Since bandwidth is a scarce resource, spectrum sensing, i.e., the identification of unused frequencies in order to opportunistically re-use these resources, will be a critical component of future wireless systems. While modern wireless systems already rely on antenna arrays to improve spectral efficiency by re-using frequencies at different locations in space, only little attention has been given to sensing and re-using unused spectrum in space. This project builds upon recent advances in low-power wideband transceiver design, wideband compressive sensing, and multi-antenna wireless communication in order to perform SPAtio-Spectral Sensing (SPASS). SPASS enables the identification of unused resources in both frequency and space, which can be used to optimize the spectral- and energy-efficiency of wireless systems. This project pursues a vertically-integrated research approach spanning circuit design, theory and algorithms, and system design in order to demonstrate the efficacy and limits of SPASS. Broader impacts resulting from this work include an extensive outreach plan involving undergraduate students from Latin America in this research and increasing participation of female high-school students with Cornell's CURIE Academy. Multi-antenna spectrum sensing via NonUniform Wavelet Sampling (NUWS), a recently introduced sampling paradigm specifically designed for RF feature extraction, is the core concept of this project. The key idea explored here is to use NUWS circuitry to extract spectral features at a coherent antenna array. These features are then used to identify unused resources in both frequency and space by means of specialized white-space detection algorithms. The project focuses on a holistic and practical approach to realizing a full and usable SPASS system; this is done by fusing the design of algorithms with the design of mixed-signal circuits for multi-antenna NUWS that extract wideband spectral features with high sensitivity directly from RF signals in an energy-efficient manner. To identify unused resources, white-space detection algorithms that exploit the sparse and low-rank structure of spectral activity in both frequency and space will be developed. The circuit prototypes and algorithms will be used to perform real-world RF measurements, which enable (i) a fundamental analysis of the system-level tradeoffs of NUWS-based SPASS and (ii) an investigation of the unused resources in realistic IoT communication scenarios that reveal the potential of reorganizing the RF spectrum in frequency and space. The circuits and algorithms to be developed find use in a broad range of other applications that rely on the extraction of spectral features from multi-channel time-domain signals. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.