Comparison of a jump-diffusion tracker with a Kalman tracker: an evaluation with emphasis on air traffic control

Report (1983)

Authors

Copyright: Campus only
Algorithms Air traffic control Clutter Diffusion theory Surveillance radar Radar tracking Aircraft manoeuvres Kalman filters Flight paths Tracking filters Standard deviation Tracking position Jump detection Monte carlo method Markov processes Turning flights
More Info
expand_more

Published Date

30-03-1983

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Source:

NLR-TR 83063 U

Abstract

A sophisticated starting point for a probabilistic approach to the radar tracking problem is a Markov jump-diffusion model for the aircraft dynamics, its control and the radar measurements. From nonlinear filtering theory, a closed form description of the evolution of the conditional distribution of this Markov process can be obtained. This jump-diffusion filter, however, is infinite dimensional and approximations are necessary for algorithmic implementation. The indirect approach of approximating the jump-diffusion by a diffusion leads to a Kalman-like tracker. The recently developed approach of approximating the jump-diffusion filter directly, leads to a bank of interacting Kalman-like trackers. The report is directed to the evaluation of tracking algorithms that are based on these two approaches. Their results are compared in an air traffic control environment. It is concluded that the jump-diffusion' tracker performs considerably better than the Kalman tracker.