2025_programme: J-divergence detection currency after thresholding a Rician signal in Gaussian noise

• Day: June 19, Thursday
    Location / Time: B. ERATO at 11:00-11:20
• Last minutes changes: -
• Session: 16. Sonar performance modeling and verification: Active and passive sonar
  Organiser(s): Mathieu Colin, Kevin Heaney
  Chairperson(s): Kevin Heaney, Victor Oppeneer
• Lecture: J-divergence detection currency after thresholding a Rician signal in Gaussian noise
  Paper ID: 2104
  Author(s): Douglas Abraham
  Presenter: Douglas Abraham
  Abstract: Sonar performance modeling traditionally entails the use of single-measurement probabilities of detection and false alarm as metrics. Extension to systems utilizing multiple measurements is often complicated enough that the scenarios are simplified to fit the single-measurement tools, thereby approximating system-level performance. By allowing an approximate representation of single-measurement detection performance in the form of J-divergence detection currency (JDC), a more accurate assessment of system-level performance can be obtained across multiple measurements. As a measure of detection performance potential, JDC can be evaluated at different points in the signal and information processing chain. After coherent detection processing, a common step in multiple-measurement systems is the thresholding of individual measurements before they are combined in a final detection decision. For example, this typically occurs when combining measurements across waveforms in active sonar or across sensors in distributed systems. The focus of this paper is on techniques for evaluating JDC after thresholding for an integrated-intensity detector operating on the standard sonar signal models (Gaussian-fluctuating, deterministic, and the more general Rician model) in Gaussian noise.
    Download the full paper
• Corresponding author: Dr Douglas Abraham
  Affiliation: Applied Physics Laboratory, University of Washington
  Country: United States