Statistical Decision Thresholds for Pulse Template Cross-Correlators

PDF

Authors
  1. Dilkes, F.A.
Corporate Authors
Defence R&D Canada - Ottawa, Ottawa ONT (CAN)
Abstract
A statistical formalism is proposed to evaluate the performance of waveform-based signal detection techniques for Electronic Support. The formalism applies to a large class of cross-correlation methods that attempt to match the arrival times of a sequence of pulses against an a priori pulse template. In this framework, the detection of a particular signal of interest is treated as a problem from binary decision theory and the observations are treated as outcomes of a point process. The presence of a signal is modeled using a pulse template with non-cumulative Gaussian jitter while background noise is modeled using a stationary Poisson point process. Two specific cross-correlation techniques are evaluated and their performances are compared in terms of false-alarm and detection probabilities. The capability of these techniques to discern between the presence and absence of the signal is measured as a function of a decision threshold.

Il y a un résumé en français ici.

Keywords
Pattern Recognition;Pulse Processing;Radar;Cross-correlation;Detection;Point process;Receiver Operating Characteristic
Report Number
DRDC-OTTAWA-TM-2005-167 — Technical Memorandum
Date of publication
01 Nov 2005
Number of Pages
58
DSTKIM No
CA026743
CANDIS No
524678
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: