Monopulse radar track‐before‐detect using Bernoulli filter

Abstract

Detection and tracking of a dim target with monopulse radar is a challenging problem. When the signal-to-noise ratio (SNR) is low, the angle estimation performance and the detection performance are poor, which makes the capture and close-loop tracking of the target very hard since the monopulse radar has a narrow beamwidth. This paper uses the track-before-detect (TBD) technique to address this problem. A measurement model is established using the Swerling II radar target model. A Bernoulli TBD filter for the monopulse radar is developed in a random finite set framework and is implemented approximately as a particle filter. The performance of the filter is studied via Monte Carlo simulations, which show that the proposed filter outperforms the classical detection and estimation methods, and it can efficiently detect and track the target at a SNR as low as −6 dB.