Torvik, B; (2016) Investigation of non-cooperative target recognition of small and slow moving air targets in modern air defence surveillance radar. Doctoral thesis , UCL (University College London).

Preview

Text Torvik_2016-12-13_PhdThesisTorvik.pdf Download (49MB) | Preview

Abstract

This thesis covers research in the field of non-cooperative target recognition given the limitations of modern air defence surveillance radars. The potential presence of low observable manned or unmanned targets within the vast surveillance volume demand highly sensitive systems. This may again introduce unwanted detections of single birds of comparable radar cross section, previously avoided by use of wide clutter rejection filters and sensitivity time control. The demand for methods effectively separating between birds and slow moving manmade targets is evident. The research questions addressed are connected to identification of characteristic features of birds and manmade targets of comparable size. Ultimately the goal has been to find methods that can utilize such features to effectively distinguish between the classes. In contrast to the vast majority of non-cooperative target recognition publications, this thesis includes non-rigid targets covering a range of dielectric properties and targets falling in the resonant and Rayleigh scattering regions. These factors combined with insufficient spatial resolution for classification require alternative approaches such as utilization of periodic RCS modulation, micro-Doppler- and polarimetric signatures. Signatures of birds and UAVs are investigated through electromagnetic prediction and radar measurements. A flexible and fully polarimetric radar capable of simultaneous operation in both L- and S-band is developed for collection of relevant signatures. Inspired by the use of polarimetric radar for classification of precipitation covered in the weather radar literature, focus has been on using similar methods to recognize signatures of rotors, propellers and bird wings. Novel micro-Doppler signatures combining polarimetric information from this sensor is found to hold information about the orientation of such target parts. This information combined with several other features is evaluated for classification. The benefit from involving polarimetric measurements is especially investigated, and is found to be highly valuable when information provided by other methods is limited.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as JSONCSVXML

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item