Coherent Multilook Radar Detection for Targets in Pareto Distributed Clutter.
Scientific Publication
- Report Number:
- DSTO-TR-2646
- Authors:
- Weinberg, G.V.
- Issue Date:
- 2012-01
- AR Number:
- AR-015-196
- Classification:
- Unclassified
- Report Type:
- Technical Report
- Division:
- Electronic Warfare & Radar Division (EWRD)
- Release Authority:
- Chief, Electronic Warfare & Radar Division
- Task Sponsor:
- DGAD
- Task Number:
- AIR 07/040
- File Number:
- N/A
- Pages:
- 20
- References:
- 23
- Terms:
- Target acquisition; Maritime surveillance; Maritime detection; Sensors
- URI:
- http://hdl.handle.net/1947/10177
Abstract
The Pareto distribution has been proposed recently as a model for intensity clutter measurements, for maritime high resolution radar returns. Using the theory of spherically invariant random processes, the Neyman-Pearson optimal detector is derived. The generalised likelihood ratio test yields a simple sub- optimal approximation, and its performance is gauged against the whitening matched filter.
Executive Summary
This work supports the Microwave Radar Branch’s research efforts in the area of high grazing angle detection of small maritime targets from an airborne surveillance platform. It contributes to the research requirements of Task 07/040 (Support to AIR 70001). Detection of targets is a major function of radar systems, and hence detector performance is an important component in the efforts of Task 07/040. Detectors are designed to find targets within specific clutter models. Recently it was found the simple Pareto Distribution fitted high grazing angle clutter quite well. This was validated using the DSTO Ingara data, collected in a trial during 2004. Hence it is necessary to investigate the design of radar detection schemes in such a clutter environment.; The theory of spherically invariant random processes (SIRPs) puts this problem into a formulation that allows the determination of detection decision rules. The Pareto distribution is put into this framework, and the Neyman-Pearson detector is specified. Due to intractability of the latter, several suboptimal approximations are considered. The major contribution of this work is to examine coherent multilook detection under the assumption of a Pareto clutter model. Although the Pareto SIRP and form of the generalised likelihood ratio test for detection has been presented already in the literature, it has not been applied directly to the multilook problem under consideration.; In addition to this, it will be shown that for Pareto models whose parameters are typical of clutter obtained from a radar operating in vertical polarisation, the whitening matched filter is an appropriate suboptimal detector. This is useful because the generalised likelihood ratio test solution depends on clutter parameters, unlike the whitening matched filter.