A Parametric Generalized Likelihood Ratio Test for Airborne Bistatic Radar Systems

Paper in proceeding, 2022

One of the main objectives of a radar system is to provide target detections. That is, from observations contaminated by receiver noise and interference determine the presence or absence of targets in the current measurements. To enable target detections, the test statistics formed by the processor is dependent on an accurate estimate of the spacetime covariance matrix to characterize the influence of thermal noise and interference on the radar signal. In a side-looking monostatic configuration, the estimate is rather straight forward as the secondary data used in the estimate can be argued to be statistically identical and independently distributed as the observation in the cell under test. However, for many other radar configurations, secondary data may suffer from angleDoppler variations over the range dimension, which introduces a non-stationary behavior in the observations. If used in a detector, such secondary data may cause significantly degraded detection performance. In this work, we propose an approach which incorporates the non-stationarities of the secondary data into the generalized likelihood ratio test. Thus, we propose a scenario and range dependent parametric model of the observed data and formulate an adaptive detector based on the generalized likelihood ratio test. The presented approach is evaluated against other state-of-the-art methods for managing target detections in the presence of non-stationary secondary data in bistatic systems. The simulations indicates that the proposed approach of imposing scenario based structure on the generalized likelihood ratio test significantly contributes to an increased performance of the target detection scheme compared to the other investigated methods.