Office of Naval Research/E21-H28.

Abstract

The research during this time period has focused on achieving more accurate realizations of the SME filter that than based on the Extended Kalman Filter (EKF). Simulation studies have shown that the linearization carried out in the EKF implementation can result in degraded performance in certain situations. To correct this, we are working on a much more accurate approximation based on the idea of sampling distribution functions. A brief description of the approach is given below. It is well known that the minimum-variance estimate of a random signal x(k) having finite second moment is equal to the conditional mean. Furthermore, when the signal state model and the measurement model are linear and the probability distributions are Gaussian, the conditional mean can be computed from a unique linear combination of the measurement data. Unfortunately, when the signal or measurement model is nonlinear, the solution to the estimation problem has proven to be much more elusive. The traditional approach has been to adapt the linear theory to nonlinear signal and measurement models through the use of the Taylor Series approximation. While this approximation simplifies the nonlinear estimation problem to the point that an implementation is realizable, it has two fundamental drawbacks:

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 30, 1994
Accession Number
ADA298914

Entities

Organizations

  • Office of Naval Research

Tags

DTIC Thesaurus Topics

  • Algorithms
  • Data Processing
  • Databases
  • Difference Equations
  • Distribution Functions
  • Euler Angles
  • Filters
  • Kalman Filters
  • Measurement
  • Military Research
  • Multitarget Tracking
  • Pattern Recognition
  • Probability
  • Probability Distributions
  • Simulations
  • Target Tracking
  • Targets

Fields of Study

  • Engineering

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Statistical inference.
  • Systems Analysis and Design