A MATH MODEL FOR COMPUTING NOISE VARIANCE MATRICES FOR A SYSTEM OF RADAR TRACKERS.

Abstract

The paper presents a method of calibrating (noisy) sensors using an unweighted least-squares procedure in a vector space setting and using orthogonal projections to implement the least-squares criterion. Using the least-squares estimate of the parameters that characterize the sensors, a variance matrix of the noise on each sensor is then computed. The method is illustrated by an analysis of both a scalar process (e.g., output of amplifiers) and a multi-variable process (e.g., output of amplifiers) and a multi-variable process (e.g., radars tracking an object). The scheme, it should be pointed out, is independent of the inputs to the system under study and hence provides a powerful tool for analysis. Finally, derivations are made in a vector-space setting employing vector-matrix techniques and using an adapted version of the notation of Dirac and Friedman. The topics covered include such items as linear transformations, vector spaces, vector 'packaging', variance analysis of vector (multi-variable) processes, least squares via orthogonal projections, and others. (Author)

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1967
Accession Number
AD0654986

Entities

People

  • Alfonso Diaz
  • James S. Pappas.

Organizations

  • United States Army Test and Evaluation Command

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Amplifiers
  • Analysis Of Variance
  • Data Science
  • Information Science
  • Mathematics
  • Notation
  • Packaging
  • Vector Spaces

Fields of Study

  • Engineering

Readers

  • Calculus or Mathematical Analysis
  • Neural Network Machine Learning.
  • Regression Analysis.

Technology Areas

  • Space
  • Space - Space Objects