Factor Analysis for Spectral Reconnaissance and Situational Understanding

Abstract

The goal of the STIR proposal is to investigate the use of tensor decompositions to perform RF tomography in a battlefield environment. Spectrum sensing techniques can be used to estimate compound spectra (comprising contributions from multiple transmitters), but identifying the constituent spectra and the locations of the ambient transmitters from the observed spectral mixtures is a challenge. The approach considered in this project is to utilize multiple linear arrays, spatial correlation lags, and frequency bins to create a tensor. Tensor decompositions will be utilized (such as PARAFAC) to locate emitters and to understand the overall spectrum utilization in the area. Research will include investigating robustness to jamming, censoring corrupt data, and potentially experimental validation.

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 12, 2017
Source ID
W911NF1510384

Entities

People

  • Nikolaos Sidiropoulos

Organizations

  • Army Contracting Command
  • United States Army
  • University of Minnesota

Tags

Fields of Study

  • Engineering

Readers

  • Radio communications and signal processing.
  • Research Science/Academic Research
  • Sensor Fusion and Tracking Systems.