Phase retrieval algorithm applied to radar characterization
Hossain, Syed Abu Sina Al
MetadataShow full item record
Phase retrieval algorithms are employed for the determination of the phase function of a complex wave from its known intensity. In this thesis, the intensity distribution considered is the modulus of the far field radiation pattern of a radar antenna. In principle, if the phase of the radiation pattern can be determined, the original near field antenna pattern can be obtained via an inverse Fourier transform. The problem to be solved here is to determine the antenna function of an aircraft radar given a noisy measurement of its far field radiation pattern. The most widely used phase retrieval algorithms use an iterative Fourier transform algorithm which involves the transformation back and forth between the Fourier domain, where the Fourier modulus data are applied, and the object domain, where the a priori object constraints are applied. These algorithms work well in the noiseless case and have been modified to handle cases with known noise statistics. In this work, the algorithms have been extended to operate in the situation where the noise statistics are not known a priori. Two different formulations are proposed and are tested on both simulated and actual aircraft radar far field antenna patterns.