Piracy, smuggling and illegal fishery threaten the overall security on oceans and seas. These threats typically arise from small and agile boats and are distributed over large areas. To control them, small maritime targets have to be detected and observed. Maritime airborne radar systems are capable of monitoring large areas and are therefore suitable to accomplish this challenge. The detection of small boats, however, is still an unresolved task due to the small radar cross section (RCS) of these boats. Additionally, the RCS of
sea clutter rises for high altitude platforms due to the higher grazing angle, hence targets with low signal-to-clutter plus noise ratios (SCNR) have to be
detected.
In order to investigate the appropriate processing to detect small boats from airborne radars, data from experiments over the North Sea with the multichannel radar system PAMIR and a small cooperative boat is evaluated in this thesis. This analysis demonstrates on one hand that traditional processing is not sufficient to detect these maritime targets, and on the other hand that with space-time adaptive processing (STAP) superior detection performance is achieved.
To apply STAP to a maritime airborne radar system, it is desirable to know its performance in advance. To accomplish this, the multichannel characteristics of sea clutter have to be understood. This thesis derives theoretical and simulation multichannel models by analyzing the spectral density matrix, the space-time filter and the channel correlation of sea clutter. Different multichannel properties for sea clutter compared to land clutter are demonstrated, which are due to the varying motion of sea echoes and due to different sea scattering types. An important implication of this difference is the broadening of the space-time filter notch in dependence on the sea condition.
To confirm the predicted properties, three measurement campaigns with the airborne radar system PAMIR were performed, where real multichannel sea data was acquired for different swell directions, two different grazing angles and significantly varying sea states. This thesis demonstrates the reproducibility of sea clutter multichannel characteristics of real data for different sea conditions with the help of calculations and simulations, which are introduced in this analysis.