An Adaptive Hull-to-Emitter Correlator

Abstract

One of the functions of a data fusion system involves identifying the platforms reported in Electronic Intelligence data. These reports contain a set of parameters measured from the platform's radar. The assignment of a name to this radar report is called Hull-to-Emitter Correlation (hultec). The authors have described an Adaptive Network System (ANS) neural network approach to this problem. The particular paradigm used was a backpropagation network. This system classifies the data into linearly separable classes, using a gradient descent, least mean squared error algorithm to determine the separating hyperplanes. The network returns a value for each hull that gives a measure of its distance from a point in the teaching set (actually it's a measure of the distance from the separating hyperplane). This approach compared favorably to traditional statistical techniques. One problem with this approach is that the output of the network, though it ranges continuously from zero to one, has no relationship to probability. Although the network gives a ranking for the hulls it has learned, the value associated with any one hull gives little information. It is only the value relative to the other hulls that can be used in further decision making. This paper deals with a modification of a statistical technique that retains the distributed nature of the ANS, while allowing the network to learn to approximate the probability distribution of the data. This not only allows the network the potential of better performance, but it allows other systems to use the output of the network in a probablistic sense.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1988
Accession Number
ADA199912

Entities

People

  • C. E. Priebe
  • D. J. Marchette

Tags

Communities of Interest

  • C4I
  • Sensors

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Bayesian Networks
  • Correlators
  • Data Fusion
  • Data Sets
  • Electronic Intelligence
  • Information Processing
  • Information Science
  • Information Systems
  • Neural Networks
  • Probability
  • Probability Density Functions
  • Probability Distributions
  • Security
  • Statistical Analysis
  • Statistics

Fields of Study

  • Computer science

Readers

  • Neural Network Machine Learning.
  • Radar Systems Engineering.
  • Statistical inference.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms
  • AI & ML - Neural Networks
  • Microelectronics