Sensor management for static target detection with non-binary sensor observations and observation uncertainty


Conference Paper

Previously, a grid-based sensor management framework has been developed that is useful for directing the operation of a suite of sensors seeking to detect static targets. Earlier versions of the framework only allow the sensors to make binary observations - either "target present" or "no target present" - within the cells of the grid. This paper introduces the use of non-binary observations within the sensor management framework. Simulation results are presented which show that the presented sensor manager continues to outperform a direct search technique with the use of non-binary observations. The effects of uncertain sensor observations are also examined in this paper, and uncertainty modeling is introduced in order to allow the sensor manager to model uncertain non-binary sensor observations. Simulation results show that proper uncertainty modeling is crucial for maintaining robust sensor manager performance. ©2007 IEEE.

Full Text

Duke Authors

Cited Authors

  • Kolba, MP; Collins, LM

Published Date

  • December 1, 2007

Published In

  • Ieee Workshop on Statistical Signal Processing Proceedings

Start / End Page

  • 74 - 78

International Standard Book Number 10 (ISBN-10)

  • 142441198X

International Standard Book Number 13 (ISBN-13)

  • 9781424411986

Digital Object Identifier (DOI)

  • 10.1109/SSP.2007.4301221

Citation Source

  • Scopus