Blunder Detection Using a Sequential Least Squares Adjustment.

Abstract

Blunder detection is a key step in achieving optimal results from a least squares adjustment. One approach to blunder detection is to employ a detection algorithm in a sequential least squares adjustment. A sequential adjustment technique compares the variability of the observations added sequentially to that of the observations already incorporated into the solution. The rejection criterion for observations is based upon an F-statistic. The F-statistic is formed by the ratio of key, independent quadratic forms of residuals corresponding to the observations added sequentially and those already accepted in the solution. Neither an a priori nor an a posteriori reference variance is required for the blunder detection algorithm. The sequential algorithm is computationally efficient because it combines the least squares adjustment and the blunder detection into one step. The sequential least squares blunder detection algorithm is programmable and is tested on measured data sampled from a simple model, a straight line. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 1984
Accession Number
ADA149514

Entities

People

  • D. A. Campbell

Organizations

  • Virginia Tech

Tags

Communities of Interest

  • C4I

DTIC Thesaurus Topics

  • Algorithms
  • Civil Engineering
  • Computer Programming
  • Computer Programs
  • Computers
  • Data Science
  • Data Sets
  • Information Science
  • Knowledge Management
  • Models
  • Notation
  • Personal Computers
  • Quality Control
  • Reliability
  • Statistical Tests
  • Surveys
  • United States

Readers

  • Computer Vision.
  • Educational Psychology
  • Operations Research