Pattern Theoretic Bayesian Inference for Multisensor Fusion.

Abstract

The focus of our Phase I study was to develop and test an algorithm to track a single aircraft of known type using the jump-diffusion method for state estimation. This technique avoids the limitations of conventional state estimation methods (most notably Kalman filtering) in dealing with the discrete state variables (such as target type) and the nonlinear/non-Gaussian measurements encountered in multisensor data fusion problems. The jump-diffusion technique operates by using Monte Carlo simulation to directly sample from the Bayes posterior distribution for the target state. The method is applied to the tracking of a maneuvering air target based on a combination of radar point tracking and optical imagery data. Numerical results show in particular that the aircraft orientation information extracted through the processing of optical images can significantly reduce tracking error relative to conventional point tracking methods. In Phase II we plan to continue the development of the Phase I algorithm by adding the capability to track multiple targets and to perform the automatic target recognition function when targets are of unknown type. (AN)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 01, 1996
Accession Number
ADA304190

Entities

People

  • Barry Belkin
  • Stephan J. Suchower

Tags

Communities of Interest

  • Air Platforms

DTIC Thesaurus Topics

  • Algorithms
  • Bayesian Inference
  • Data Fusion
  • Data Science
  • Information Processing
  • Information Science
  • Kalman Filtering
  • Monte Carlo Method
  • Multiple Targets
  • Optical Images
  • Recognition
  • Statistical Algorithms
  • Target Recognition
  • Targets

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Mathematical Modeling and Probability Theory.
  • Radar Systems Engineering.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms