Choosing Candidate Locations for Source Localization

Abstract

Several localization algorithms calculate the location by performing a search of the space of possible source locations. Included among these algorithms is ML [1] and SRP-PHAT [2]. The source is assumed to be in a finite space known as the region of interest. The estimated location is the candidate location with the greatest energy. If further refinement is desired, the resulting estimate can be used as an initial location for a hill-climbing algorithm [1]. However, one issue that has not been addressed in the literature is how to efficiently choose the candidate locations in order to best represent the space. If too many candidates are used, the computational complexity is greater than needed. If too few candidates are used the space will not be accurately described and the estimation accuracy will suffer. A general method for determining the optimal spacing of candidate locations will be presented. This method is based on ideas from image warping [3].

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 2005
Accession Number
AD1170834

Entities

People

  • Chris Kyriakakis
  • J. M. Peterson

Organizations

  • University of Southern California

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • California
  • Climbing
  • Computational Complexity
  • Digital Images
  • Electrical Engineering
  • Engineering
  • Errors
  • Filters
  • Frequency
  • Images
  • Literature
  • Low Pass Filters
  • Microphones
  • Near Field
  • Sampling
  • Signal Processing

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Seismology

Technology Areas

  • Space