Techniques for the Statistical Analysis of Observer Data

Abstract

For vehicle designers, the main goal of experiments on the observability of combat vehicles is a comparison of the probability of detection of two vehicles as a function of range. This paper investigates two statistical techniques for analyzing data from fixed observer tests. The two techniques are as follows: (1) fitting logistic curves to the vehicle data, and (2) using the Fisher Exact Test to compare the probability of detection of the two vehicles at each range. The paper also discusses the issues of background variability and confidence levels for hypothesis testing. Results show that the Fisher Exact Test has advantages over fitting a logistic curve. Because data are compared only at the same range, the effect of variability of background with range is avoided. Experimental sample sizes can be calculated for comparison of proportions to insure that a given difference in probability of detection (Pd) will be significant. Moreover, the Fisher Exact Test lends itself to quantitative hypothesis testing.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 14, 2001
Accession Number
ADA460096

Entities

People

  • John G. Bennett

Organizations

  • Tank-automotive and Armaments Command

Tags

DTIC Thesaurus Topics

  • Abstracts
  • Combat Vehicles
  • Detection
  • Information Operations
  • Mathematics
  • Observers
  • Probability
  • Statistical Analysis
  • Vehicles

Fields of Study

  • Mathematics

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Aerospace Test and Evaluation
  • Computational Modeling and Simulation