Classification by Markovian Strings of Pings,

Abstract

The probability of correct target identification is calculated for sequences of pings. The inputs to the calculation are single hit and false alarm probabilities as well as transition probabilities. The inclusion of transition probabilities allows for the possibility that the decision with regard to a given echo may effect the decision about subsequent echoes. Probability of correct target identification for sequences of n pings is calculated by using the maximum likelihood criterion. Numerical results are shown that indicate excellent agreement with the empirical findings of Stradling and Schumacher in connection with the TRESI evaluation. To achieve this agreement it is necessary however to take advantage of the Markovian assumption since in the special case of independence the results of the calculations do not agree with experiment. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 18, 1966
Accession Number
ADA071227

Entities

People

  • Henry P. Kramer

Organizations

  • General Electric

Tags

Communities of Interest

  • Human Systems
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Agreements
  • Classification
  • Computer Programs
  • Detection
  • False Alarms
  • Hit Probabilities
  • Identification
  • Markov Chains
  • Markov Processes
  • Probability
  • Random Variables
  • Recognition
  • Sequences
  • Sonar Echoes
  • Transitions
  • Warning Systems

Readers

  • Acoustical Oceanography.
  • Mathematical Modeling and Probability Theory.
  • Statistical inference.