Robust Multi Sensor Classification via Jointly Sparse Representation

Abstract

In this project, we have developed various novel collaborative sparse representation methods for multi-sensor classification problem, which take into account correlation as well as complementary information between heterogeneous sensors simultaneously while considering joint sparsity within each sensor's observations. We also robustify our models to deal with the presence of sparse noise and low-rank interference terms. Especially, we observe that incorporating the noise or interfered signal as a low-rank component is essential in a multi-sensor problem when multiple co-located sources/sensors simultaneously record the same physical event. Essentially, our proposal combines the strengths of multiple ideas: (i) incorporating related information from different sources (sensors) to achieve an improvement in the classification performance; (ii) extracting and suppressing a large, dense and correlated (hence low-rank) signal/noise interference normally appeared in multi-sensor data; and (iii) exploiting prior structure in sparsity representations for efficiency and robustness.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 14, 2016
Accession Number
AD1009706

Entities

People

  • Trac D. Tran

Organizations

  • Johns Hopkins University

Tags

Communities of Interest

  • Sensors

DTIC Thesaurus Topics

  • Algorithms
  • Compressed Sensing
  • Computer Programs
  • Data Sets
  • Detection
  • Detectors
  • Earth Sciences
  • Engineering
  • Image Classification
  • Image Processing
  • Information Processing
  • Information Theory
  • Kernel Functions
  • Remote Sensing
  • Signal Processing
  • Students
  • Supervised Machine Learning

Readers

  • Computer Vision.
  • Distributed Systems and Data Platform Development
  • Operations Research