Alternatives to an Extended Kalman Filter for Target Image Tracking.

Abstract

Four alternative filters are compared to an extended Kalman filter (EKF) algorithm for tracking a distributed (elliptical) source target in a closed loop tracking problem, using outputs from a forward looking (FLIR) sensor as measurements. These were (1) an EKF with (second order) bias correction term, (2) a constant gain EKF, (3) a constant gain EKF with bias correction term, and (4) a statistically linearized filter. Estimates are made of both actual target motion and of apparent motion due to atmospheric jitter. These alternative designs are considered specifically to address some of the significant biases exhibited by an EKF due to initial acquisition difficulties, unmodelled maneuvering by the target, low signal-to-noise ratio, and real world conditions varying significantly from those assumed in the filter design (robustness). Filter performance was determined with a Monte Carlo study under both ideal and non ideal conditions for tracking targets on a constant velocity cross range path, and during constant acceleration turns of 5G, 10G, and 20G.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1981
Accession Number
ADA111144

Entities

People

  • Paul R. Leuthauser

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • C4I
  • Sensors
  • Weapons Technologies

DTIC Thesaurus Topics

  • Acquisition
  • Air Force
  • Algorithms
  • Computational Science
  • Computations
  • Estimators
  • Filters
  • Kalman Filters
  • Mathematical Filters
  • Measurement
  • Monte Carlo Method
  • Optimal Estimators
  • Plastic Explosives
  • Probability
  • Rdx
  • Statistical Algorithms
  • Stochastic Processes

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Electromagnetic Wave Scattering and Antenna Radiation Engineering

Technology Areas

  • 5G
  • 5G - Internet of Things