Parameters Extraction of Unknown Radar Signals Using Change Point Detection

Abstract

Many applications like seismic data processing or navigation system monitoring require detection of abrupt changes. This paper considers the problem of signal detection and recognition in the context of radar application applied to a segmentation of an unknown received signal. Therefore a well known algorithm in change point detection is being implemented, the Cumulative Sum algorithm (CuSum), in order to detect and extract signals of interest in radar recording. It enables the localization of the beginning and ending of unknown waveforms in a pulse train making its extraction from a noisy background and its characterisation possible. Different models of the problems are considered and compared. The capacity of the algorithm to correctly retrieve the temporal parameters of the signal is tested on both simulations and acquisitions.