An Artificial Neural Network Tracking Architecture

Abstract

In this paper, a neural extended Kalman filter algorithm is used for tracking of a highly maneuvering target. The neural extended Kalman filter is used to improve a mathematical motion model for use in prediction. Instead of just applying a high process noise model (a catch-all technique) in an interacting multiple model architecture to hold a target through a maneuver, a neural extended Kalman filter is used to predict the correct velocity and acceleration states of a target. This, in turn, may allow noise reduction during a target maneuver. Tracking results that stress the algorithm during severe maneuvers are shown along with the tuning parameter issues of the artificial neural network.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 14, 2002
Accession Number
ADA506947

Entities

People

  • Allen R. Stubberud
  • Mark W. Owen

Organizations

  • Naval Information Warfare Systems Command

Tags

Communities of Interest

  • Materials and Manufacturing Processes
  • Sensors

DTIC Thesaurus Topics

  • Algorithms
  • Computing System Architectures
  • Data Association
  • Estimators
  • Filters
  • Kalman Filters
  • Linear Systems
  • Maneuvers
  • Military Research
  • Multitarget Tracking
  • Neural Networks
  • Noise
  • Nonlinear Systems
  • Statistical Algorithms
  • Target Tracking
  • Targets
  • United States Government

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms
  • AI & ML - Neural Networks