Estimation of Receiver Operating Characteristic Curves for Multiping Mine Detection Using Empirical Target Data

Abstract

This report develops the formalism for computing receiver operating characteristic (ROC) curves for an active sonar detector, consisting of linear receiver processing followed by a linear envelope detector. The formalism is applicable to both single-ping and multiple-ping detection. The methodology for computing the ROC curves directly incorporates measurements of target strength as a function of aspect angle to estimate the requisite signal-plus-noise probability density function (pdf) under the assumption of Gaussian background statistics. This "empirical" approach for computing the probability of detection allows ROC curve performance to be estimated for multiple ping integration and for any sensor-target scenario that is supported by the available target data.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 14, 2002
Accession Number
ADA421690

Entities

People

  • Petre I. Rusu
  • S. P. Pitt

Organizations

  • University of Texas at Austin

Tags

Communities of Interest

  • Energy and Power Technologies
  • Sensors

DTIC Thesaurus Topics

  • Aspect Angle
  • Contracts
  • Data Science
  • Detection
  • Detectors
  • False Alarms
  • Gaussian Noise
  • Information Science
  • Measurement
  • Navy
  • Probability
  • Probability Density Functions
  • Random Variables
  • Statistics
  • Target Strength
  • Targets
  • Warning Systems

Fields of Study

  • Physics

Readers

  • Acoustical Oceanography.
  • Sensor Fusion and Tracking Systems.
  • Statistical inference.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference