Radar target recognition using low frequency resonance techniques

Allan, Robert Andrew (1979) Radar target recognition using low frequency resonance techniques. University of Southampton, Doctoral Thesis.

Record type: Thesis (Doctoral)

Abstract

This thesis examines the problem of radar target recognition. The work has been primarily concerned with low frequency techniques using the resonance region. In this scattering region, the wavelength of the radar excitation is approximately the dimensions of the target. The response is largely governed by the overall shape of the scatterer and any fine details are ignored. The research in this report can he divided into two principal areas namely the calculation of resonances and the development of recognition algorithms. Five methods of calculating resonances are examined. These are the matrix methods of Prony, Mullis and Sanathanan and two serial approaches developed in the project. The prime considerations when choosing a method suitable for the radar application are speed of operation and accuracy in the presence of noisy data. Two recognition algorithms based on resonance information are described in the thesis and cover the problems of target identification and recognition. The identification alqorithm is based on a similarity measure which calculates the resemblance between two sets of resonance data. The similarity measure is derived from a modified form of Bayes Rule. The recognition algorithm associates the resonances of the target with predetermined groups of resonance locations. A method in which the optimal combination of groups can be found from the resonances of all possible targets is also described. Computer generated wire model data is used to test the identification algorithm althouqh the project terminated before any adequate testing was possible. However the preliminary results obtained are encouraging and would suggest that further research would prove worthwhile.

This record has no associated files available for download.