Migrating Target Detection in Wideband Radars

Research output: ThesisDissertation (TU Delft)

125 Downloads (Pure)

Abstract

Modern surveillance radars are designed to detect moving targets of interest in an adverse environment, which can encompass strong unwanted reflections from ground or sea surface, clouds, precipitation, etc. Detection of weak and small moving targets in environmental clutter remains, however, a challenging task for the existing radar systems. One of the main directions for modern radar performance improvement is the application of wideband high-resolution waveforms, which provide detailed range information of objects in the observed scene. Together with such inherent advantages of wideband waveforms as multi-path separation, clutter reduction and improved target classification, additional benefits can be obtained by exploiting target range migration (range walk), essential for fast moving targets in the high-resolution mode. This thesis aims at the development of novel signal processing techniques for migrating target detection in wideband radars. It involves both resolving range-velocity ambiguities and improvement in target discrimination from ground clutter by accounting for target range migration. It is demonstrated that wideband radars can resolve range-velocity ambiguity by transmitting a single long pulse burst with low pulse repetition frequency (PRF) and exploring target range walk phenomena during the burst. The ambiguity function of such waveform still has strong residuals at the locations of ambiguities, called ambiguous sidelobes, which have to be considered in the processing of wideband data. The presence of ground clutter in the observation scene has a detrimental effect on the wideband radar performance. The impact of the clutter Doppler spectrum and waveform parameters on target detection at clutter ambiguities has been investigated. The improvement over the conventional waveform is demonstrated for narrow clutter Doppler spectrum; in the presence of clutter with a wide Doppler spectrum, the conventional staggered-PRF waveform is preferable. Performance degradation at ambiguous-to-clutter velocities is validated on the real data sets. Modern high-resolution parametric-free spectrum estimators – IAA (Iterative Adaptive Approach) and SPICE (Semi-Parametric Iterative Covariance-based Estimator) – are proposed for the reconstruction of the observed scene from wideband radar measurements with no velocity ambiguities. These algorithms demonstrate significant improvement in rejection of ambiguous sidelobes over the conventional techniques. For clutter-limited case, the covariance-aware SPICE is introduced with improved capability to discriminate targets from clutter. The advantages of the proposed methods are demonstrated in numerical simulations and real data processing. The ambiguous sidelobes can cause severe problems for detection of multiple targets located at similar ranges. A dedicated detector for a dense target scenario has been introduced. It can detect multiple closely spaced targets and mitigate false detections due to their ambiguous sidelobes, holding false alarm probability at the required level. The improvement over conventional processing is demonstrated. Special attention is then devoted to clutter suppression in the high range resolution mode. In meter or sub-meter range resolution, the observed ground clutter, modeled by a compound-Gaussian process, may have significant fluctuations over the range interval, elapsed by the target. An advanced detector for range-migrating targets in compound-Gaussian clutter is developed. It performs two-dimensional clutter filtering – in Doppler frequency and in the range – and benefits from clutter spatial diversity, obtained for a target passing over different patches of clutter. A significant improvement in the detection of fast moving targets in spiky clutter is achieved in comparison to the existing methods. The attained gain depends on clutter characteristics and target velocity: fast moving targets are easier to detect than slow ones with equal signal-to-clutter ratio. The generalized approach for detection of range-extended migrating targets is provided. The performed research provides some fundamental insight for implementation of new radar architectures with the utilization of wideband waveforms.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Delft University of Technology
Supervisors/Advisors
  • le Chevalier, F., Supervisor
  • Yarovyi, O., Supervisor
Award date21 Jun 2019
Print ISBNs978-94-028-1578-8
DOIs
Publication statusPublished - 2019

Fingerprint Dive into the research topics of 'Migrating Target Detection in Wideband Radars'. Together they form a unique fingerprint.

  • Cite this