Use this URL to cite or link to this record in EThOS:
Title: Processing techniques for improved radar detection in spiky clutter
Author: Armstrong, Brian Clement
ISNI:       0000 0001 3427 7516
Awarding Body: University of London
Current Institution: University College London (University of London)
Date of Award: 1992
Availability of Full Text:
Access from EThOS:
Access from Institution:
The problem of improved radar detection of targets embedded in spiky clutter is addressed. Two main areas where improvements may be possible are investigated, namely improved clutter suppression by doppler filtering, and improved Constant False Alarm Rate (CFAR) processing. The clutter suppression performance of several doppler processors is quantified under a wide range of conditions. It is shown that in spatially homogeneous clutter ideal optimal (Hsiao) filters offer 2 to 3 dB higher improvement factor than conventional techniques. Adaptive Hsiao filters are evaluated under conditions of spatially heterogeneous clutter, and it is shown that practical losses due to filter adaptivity and spectral heterogeneity will outweigh the superior performance of ideal Hsiao filters in homogeneous clutter. It is concluded that improved doppler filtering offers little scope for improving detection performance in spiky clutter, and that more significant benefits are to be gained through improved CFAR processing. The performance of three current generation CFAR processors is evaluated in spatially uncorrelated K-distributed clutter to quantify detection losses. It is shown that losses of in excess of 10 dB can be expected in spiky clutter. Reducing the loss by exploitation of any spatial correlation of the underlying clutter power is investigated. To this end a mathematically rigorous model for spatially correlated K-distributed clutter is derived. An improved CFAR processor based on optimal weighting of reference cells is formulated and evaluated. It is shown that in highly correlated clutter CFAR loss can be reduced by 2 to 5 dB compared to Cell Averaging CFAR processors. An alternative "RDT-CFAR" processor is formulated to eliminate reliance on spatial correlation, and this is shown to reduce CFAR loss by more than 10 dB in spectrally homogeneous spiky clutter. However, an increase in false alarm rate in clutter without constant spectrum is demonstrated. The RDT-CFAR processor has been modified to eliminate dependence on surrounding range bins. The resulting "δ-CFAR" processor reduces CFAR loss by more than 10 dB in even moderately spiky clutter. It is also immune to extraneous targets and clutter edges, and its false alarm performance is insensitive to clutter spikiness.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Radar detection