Search

The optimal use of a surveillance radar system requires proper understanding about the system behavior in different configurations, modes, and operating conditions. This paper proposes a machine learning framework for producing and validating the performance model of the surveillance radar system. The framework consists of an optimization method for the parameterization of a radar model and a machine learning method for the modeling of a tracker. Optimization and machine learning is based on the satellite navigation data of cooperative aircraft and corresponding track data from the surveillance system. The aim is to learn the system performance in a wide range of operating conditions using the extensive measurement history and then to predict the present performance with high accuracy at specified locations in the airspace. The feasibility of the proposed framework is assessed using real data. acceptedVersion

The objective of the radar performance prediction is to compute quantities of interest concerning the ability of the radar to observe its surroundings. Besides the properties of the radar system, the performance is affected by the target, whose radar cross section (RCS) is one of the predominant factors. The performance prediction is usually performed in relation to the target RCS characterized by a constant value or a particular statistical distribution. Such representations generalize real-life complex targets rendering them unsuitable for some objectives since the RCS is significantly influenced by the target aspect angle and is inherently stochastic by nature. Thus, a more dynamic description may be valuable e.g. for analyzing the radar performance on a flight path of interest. We propose representing the RCS with a histogram that includes such dynamic properties and is suitable for considering the target in different ways for performance prediction: in a more general manner or dependent on its aspect angle. We consider the case of traditional RCS with low spatial resolution and demonstrate the proposed approach through the probability of detection computed for a generic surveillance radar. acceptedVersion

The radar cross section (RCS) of a target describes its reflectivity towards the radar. It is a deterministic quantity—specific to the aspect angle of the target and the properties of the radar signal—and as such can be predicted with various methods when provided with an adequate physical model of the target. However, stochasticity is introduced when considering a target during flight and several factors increasing the uncertainty—typically about the location and orientation of the target in relation to the observing radar—have to be taken into account. It is possible to model these uncertainties when adequate information about physical properties of the target as well as about its flight path and motion mode is available. In this paper, we propose a method for incorporating such knowledge about the flight dynamics into the estimation of a momentary distribution for the RCS. We use a simulated trajectory to demonstrate the method and present comparisons to the Swerling I model.

Geographical information systems (GIS) have been the base for radar ground echo simulations for many years. Along with digital elevation model (DEM), present GIS contain characteristics of terrain. This paper proposes a computationally sensible simulation procedure to produce realistic radar terrain signatures in a form of raw data of airborne pulse Doppler radar. For backscattering simulation, the model of the ground is based on DEM and built with point-form backscattering objects. In addition to the usual DEM utilization for xyz coordinates and shadowed region calculation, we assume that each data point in GIS describes several scatterers in reality. Approaching the ground truth, we distribute individual scatterers with adjustable attributes to produce authentic response of areas such as sea, fields, forests, and built-up areas. This paper illustrates the approach through an airborne side-looking synthetic aperture radar (SAR) simulation. The results prove the enhanced fidelity with realistic SAR image features.

Radars are used for various purposes, and we need flexible methods to explain radar response phenomena. In general, modeling radar response and backscatterers can help in data analysis by providing possible explanations for measured echoes. However, extracting exact physical parameters of a real world scene from radar measurements is an ill-posed problem. Our study aims to enhance radar signal interpretation and further to develop data classification methods. In this paper, we introduce an approach for finding physically sensible explanations for response phenomena during a long illumination. The proposed procedure uses our comprehensive response model to decompose measured radar echoes. The model incorporates both a radar model and a backscatterer model. The procedure adapts the backscatterer model parameters to catch and reproduce a measured Doppler spectrum and its dynamics at a particular range and angle. A filter bank and a set of features are used to characterize these response properties. The procedure defines a number of point-scatterers for each frequency band of the measured Doppler spectrum. Using the same features calculated from simulated response, it then matches the parameters-the number of individual backscatterers, their radar cross sections and velocities-to joint Doppler and amplitude behavior of the measurement. Hence we decompose the response toward its origin. The procedure is scalable and can be applied to adapt the model to various other features as well, even those of more complex backscatterers. Performance of the procedure is demonstrated with radar measurements on controlled arrangement of backscatterers with a variety of motion states.