KBARET

KBaReT (Knowledge-Based Radar for Target Detection and Identification)

The aim of this study is to present a state of the art  assessment of KBR signal and data processing techniques together with a roadmap for the implementation of this capability in European Radar  and Multifunction RF Systems . This study is envisaged to find ways of mitigating the problem of clutter returns by using smarter radar signal processing in the form of KBR processing, including  self-learning and adaptive processing techniques.

The  KbaReT project has been articulated on six WPs:

  • WP1 – State of the art (SoA) of KBR and architecture assessment.
  • WP2 – KBR Impact on Emerging Technologies.
  • WP3 -KBR for CFAR and STAP.
  • WP4 – KBR for Multifunction Sensors.
  • WP5 – KBR for Non Cooperative Identification (NCI).
  • WP6 – KBR Implementation Roadmap. A detailed description of each WP is provided in the following sub-sections.

CNIT-RaSS was involved in WP1, WP2, WP3, WP5 and WP6.

In particular, the CNIT-RaSS has performed the following technical activities:

  • Design and validation of KBR signal processing algorithms aimed at enhancing the target detection performance. KB-CFAR and  KB-STAP algorithms have been proposed and tested by means of simulated data. KB-STAP has been formulated for both a monostatic and a bistatic radar geometry. Moreover two different Knowledge based algorithm has been implemented, namely the ITFS (Intelligent Training and Filter Selection) approach and the Bayesian approach.  The main results of this activity are presented in the figures below.
  • Design and validation of a KB algorithm exploiting the polarimetry response of a target to enhance its detection in strong clutter conditions. Measures was used in this case to test and validate the developed algorithms.

Design and validation of a KB Bayesian classifier. Validation has been performed by using measures acquired in a controlled-type measurement campaigns.

PUBLICATIONS