Defense

Modeling of the water echo for multi-channel signal processing with maritime radar systems

Experiment über der Nordsee.
© Photo Fraunhofer FHR

Experiment over the North Sea.

Normierter STAP-Filtergewinn von realen (oben) und simulierten (unten) mehrkanaligen Landdaten.
© Photo Fraunhofer FHR

Normalized STAP filter gain from real (above) and simulated (below) multi-channel land data.

Normierter STAP-Filtergewinn von realen (oben) und simulierten (unten) mehrkanaligen Seedaten.
© Photo Fraunhofer FHR

Normalized STAP filter gain from real (above) and simulated (below) multi-channel sea data.

Advanced multi-channel signal processing is essential for the detection of moving targets with maritime radar systems. Here, the detection capacity can only be calculated if the multi-channel properties of the water echo are known. Fraunhofer FHR developed such a model and validated it with real data.

Maritime radar systems – from the past and present

The first radar, which was demonstrated by Christian Hülsmeyer in 1904 on the Hohenzollern bridge in Cologne, was designed to detect ships on the Rhine. Radar systems on stationary or shipboard platforms capable of detecting large ships without complicated signal processing are still of significance today. But there is also a great deal of interest in the detection of small and agile boats from airborne platforms. This would facilitate the monitoring of threats, e.g. from piracy or illegal fishing, over wide areas. Small boats, however, only return a low signal which has to be detected from a strong water echo and this gives rise to the need for advanced signal processing.

Precalculation rather that later regrets

Fraunhofer FHR demonstrated that Space-Time Adaptive Processing (STAP) is necessary for the reliable detection of small boats. STAP uses the fact that disturbing echoes exhibit a definite relationship between the direction of incidence and the radial velocity to estimate an improved filter for the suppression of this backscatter. It is important that the theoretical detection capacity of a specific radar system is known to facilitate, for example, the calculation of the optimal radar parameters or to determine whether or not the additional costs for a multi-channel system as opposed to a single-channel system are meaningful.

The capacity of STAP can only be calculated if the multi-channel properties of the disturbing echo, i.e. the clutter, are known. The multi-channel properties of land clutter have been recorded but these, however, are largely static and are therefore not suitable for maritime radar systems. Sea clutter is generated from a moving surface and scatter from breaking waves also has to be considered.

At Fraunhofer FHR, a model was developed for the simulation and calculation of the multi-channel water echo and the associated important parameters for STAP. Several measurement campaigns with the multi-function radar system PAMIR were carried out to validate the model. In this context, the carrier platform Transall C-160 flew over the North Sea near Helgoland, as shown in Figure 1.

Land echo model not suitable for water

When determining the detection capacity it is essential that the STAP filter is known as this not only eliminates cluster but may also attenuate or even suppress the target signal. This filter, which is estimated adaptively from the data, indicates the extent to which a target is attenuated depending on the direction cosine and the radial velocity. Figure 2 shows the normalized gain of such a filter from a real and a simulated data set. One can see that this filter corresponds to a diagonal, i.e. the target is only attenuated for a specific radial velocity in each line of vision.

Figure 3 shows the normalized filter gain of a real and a simulated data set. This filter distinguishes itself significantly from the filter in Figure 2 through the wider and asymmetrical notch. The broadening of the filter is attributable to the movement of the water and the asymmetry is caused by the scatter from breaking waves. The detection of a target with a low signal is more difficult with this filter, in particular for negative radial velocities. The real filter of the sea data set shows that the land clutter model can not be used to determine the detection capacity on water as this would lead to incorrect results. The filter gain of the simulated data shows that the developed model reproduces the real multi-channel properties in a satisfactory manner and is therefore suitable for the determination of the detection capacity of potential maritime targets.