×

Toeplitz Hermitian positive definite matrix machine learning based on Fisher metric. (English) Zbl 1458.94078

Nielsen, Frank (ed.) et al., Geometric science of information. 4th international conference, GSI 2019, Toulouse, France, August 27–29, 2019. Proceedings. Cham: Springer. Lect. Notes Comput. Sci. 11712, 261-270 (2019).
Summary: Here we propose a method to classify radar clutter from radar data using an unsupervised classification algorithm. The data will be represented by positive definite Hermitian Toeplitz matrices and clustered using the Fisher metric. Once the clustering algorithm dispose of a large radar database, new radars will be able to use the experience of other radars, which will improve their performances: learning radar clutter can be used to fix some false alarm rate created by strong echoes coming from hail, rain, waves, mountains, cities; it will also improve the detectability of slow moving targets, like drones, which can be hidden in the clutter, flying close to the landform.
For the entire collection see [Zbl 1428.94016].

MSC:

94A12 Signal theory (characterization, reconstruction, filtering, etc.)
15B05 Toeplitz, Cauchy, and related matrices
62H30 Classification and discrimination; cluster analysis (statistical aspects)
PDFBibTeX XMLCite
Full Text: DOI HAL