A Generic Phased Array Radar Model for detailed radar performance assessment

Abstract

The Defence Science and Technology Organisation of Australia undertakes the role of evaluating proposals for the acquisition and upgrade of military systems. Simulation is increasingly employed as a means of assessing military effectiveness and systems performance. This might be because the systems concerned exist only on paper or because there are too many variables affecting the system's performance that experimental trials would be prohibitively expensive on their own. In the domain of microwave radar, the Electronic Warfare & Radar Division of DSTO has developed a detailed simulation model called GPARM (Generic Phased Array Radar Model) which represents the radio-frequency environment, the radar's antenna, and the radar's array and signal processing. The radio-frequency environment module of GPARM, in particular, is very detailed and represents the returns of radar signals from land and sea clutter, targets, emitters and the modulation of signals due to anomalous propagation effects. The model is fully three-dimensional in order to capture the effects of clutter in the antenna sidelobes. This is particularly important for modelling modern digital phased array antennas which employ adaptive beamforming on receive to minimise the effects of spatially-distributed clutter and to spatially cancel deliberate jamming, because the receive weights used are a function of the signals returned from the three dimensional radio-frequency environment. Conventional two-dimensional models which represent a vertical slice through the main beam are unable to capture this aspect of a radar's performance. The simulation model is designed to be modular so that the domain expert in a particular aspect of the system and environment model, such as statistical clutter modelling or array signal processing, can contribute or modify a module of interest without having to comprehend every other aspect of the model's operation. These modules could collectively be regarded as populating a toolbox but as GPARM is a time-based simulation there needs to be an overall infrastructure to manage the time-stepping. Whereas the most commonly-used radar system performance models, which are said to be parametric in the sense that they represent the effect of physical and algorithmic processes on the target signal to noise or clutter mean power ratio, GPARM works directly with the simulated time-sampled signals which originate from sampled probability density functions representing clutter and noise. Because of this, GPARM can explicitly process the sampled signals using the same signal and array processing algorithms which would be used in an operational radar, rather than use the approximate models for these processes used by parametric models, which usually make assumptions of Gaussian statistics. However, since the signals represent statistical samples of random processes, average performance has to be estimated using Monte Carlo simulation. Rather than explicitly represent the transmitted waveform at its carrier frequency, the effect of the attenuation of the signal by the environment in terms of amplitude and phase is represented at base-band, that is, as if the carrier had been down-converted to 0 Hz. This enables the signals to be represented at the sample rate of the radar. The purposes to which this model may be put are several-fold. It may be used to assess the performance of a radar system within a complex operational scenario which would be difficult to arrange in practice; to provide simulated data for the development and testing of signal processing, detection and tracking algorithms; for the off-line processing of measured data; and as a design tool for exploring trade-offs in the design of the overall system.