Research On Phaseless Near Field Antenna Measurement

The phaseless planar near-field antenna measurement is a key technology in antenna measurement.On the one hand,with the fast development of communication and detection technology,the operating frequency of electronic equipment is growing higher and higher,and even the millimeter waves is widespread.Then,it is difficult to measure the phase of the electric field with a vector network analyzer with high accuracy.On the other hand,in some special measurement scenarios,a vector network analyzer cannot be used,but only a spectrum analyzer can be used for measurement,which cannot acquire the phase of the electric field.In the above two cases,only the amplitude of the electric field can be measured.Therefore,two planes are often used for scanning the amplitudes of the electric field,which will be applied to compute the phases of the electric field by the phase recovery algorithms.The phase retrieval problem is an NP hard mathematical problem,and it is originated from optical imaging.With a long term of research effort over nearly 100 years,dozens of different phase retrieval algorithms have been proposed.These algorithms have gradually matured and have been gradually applied to image processing,computational biology,and blind deconvolution,etc.And they have achieved many great effects.However,only a few of them are applied to electromagnetics.This thesis attempts to use the existing phase retrieval algorithms to implement the phaseless planar near-field antenna measurements.Most of these phase retrieval algorithms are iterative methods based on numerical optimization,which are mainly divided into two categories: the convex optimization methods and the non-convex optimization methods.At first,this thesis details the theory of antenna near-field measurement,especially the integral equation of electric field in terms of the plane wave spectrum.Then it focuses on the phase retrieval method based on non-convex optimization.Finally,the integral equation of the electric field is converted into a matrix equation,which is solved by optimization to restore the phase distribution.The main research work and innovations are as follows:(a)The basic theory of plane near-field measurement is studied,and the plane near field to far field transform(NF-FF)algorithm is introduced.The accuracy of the algorithm is validated by the virtual measurement example of a horn antenna and a Taylor antenna array.Given the electric field on a scanning surface,the near-field distribution on another plane and the far-field pattern obtained by the introduced algorithm are in good agreement with the results of FEKO.(b)The probe compensation theory based on Lorentz reciprocity theorem is studied,and the probe compensation algorithm for the NF-FF is deduced.An example of measuring a Taylor array with a Yagi antenna shows that the far-field phase pattern and amplitude pattern calculated by this probe compensation algorithm are correct.(c)Various phase retrieval algorithms based on numerical optimization are studied,such as Gerchberg-Saxton method(GS),Fienup algorithm,Wirtinger flow based method(WF),truncated Wirtinger flow based method(TWF)and truncated amplitude flow based method(TAF),etc.Since these algorithms are tailored for problems arising from various scientific fields,they cannot be easily applied to electromagnetics.Therefore,it is necessary to test their accuracy,convergence rate and time expenditure with several simple matrix equations.(d)Various phase retrieval algorithms,especially GS,Fienup and TAF,are applied to the phaseless near-field antenna measurement.It is found that TAF has the fast convergence rate.Because it employs a random searching direction to get rid of the local minimum.Theoretically speaking,with sufficient number of iterations,TAF could recover the solution exactly with high probability and linear complexity.