Research On Time Reversal Based Target Imaging Techniques

Time reversal(TR)is a new signal processing technique,it exploits the invariance of the wave equation in lossless and time-invariant media,and involves physical or synthetic retransmission of signals recorded by multiple sensors in a time reversed fashion to focus at the original signal source or target position.In electromagnetic imaging,the most essential difference between time reversal and traditional imaging techniques is that TR effectively combines the physical transmission of waves and array signal processing,and can achieve synchronously spatial-temporal focusing which is self-adaptive to the environment.Especially in the complex propagating background TR can expand the equivalent aperture of regular array by exploiting the multipath information,which improves the system imaging performance and environmental adaptability.These unique properties make TR has a wide application prospect in target detection,identification,location and tracking.Meanwhile,lots of aspects including imaging resolution,stability and technical extension of some existing methods are still to be improved.Therefore,researching time reversal imaging technique has vital theoretical significance and application value.In this dissertation,the mechanism,characteristics and shortcomings of existing methods such as the time reversal mirror(TRM)and time reversal transmission matrix are deeply analyzed based on the fundamental theory of TR.Three new imaging methods are proposed to make improvement and innovation.Furthermore,based on the study of frequency diverse array(FDA)technique,the TR is combined with it,and a way to enhance the target parameters estimation performance of existing FDA radar is proposed.The main work of the dissertation are summarized as follows:At first,TR basic theories and methods are studied.The fundamental principle and spatial-temporal synchronous focusing characteristic of TR are analyzed,based on which several major TR imaging approaches such as TRM,time reversal operator(TRO)decomposition and space-frequency TR are discussed in detail.This lays a theoretical foundation for the following research.Secondly,target imaging based on TRM is investigated.The TRM target detecting model is built by finite-difference time-domain method,several algorithm for solving the TRM optimal time frame(OTF)are evaluated,and a modified Multi-TRM method is proposed for the near-far problem of traditional TRM in multi-target imaging.It is based on the time domain gating,grouping isolation and time-sharing retransmission of echo pulses corresponding to different scatterers,and outputs the TR field revised by zero setting,superposition and normalization process to obtain the imaging results.Simulation results show that the new method eliminates the interferences among retransmitted signals of different scatterers,and can image all targets in multi-target scenario,which promotes the effectiveness of TRM for multi-target detection.Next,target imaging based on TR transmission matrix is investigated.Based on the analysis of space-frequency TR,the estimation of signal parameters via rotation invariance techniques(ESPRIT)model is introduced,and the SF-ESPRIT single point target imaging method is proposed.Compared to the conventional TRO decomposition method,the new method requires only single transceiving of antenna array to acquire echoes to build SF-MDM.After performing singular value decomposition(SVD),the target pseudo-spectrum is constructed based on the displacement invariance of subarrays and rotation invariance of signal subspace to obtain the imaging results.Simulation results show that under good noise conditions the new method has an imaging resolution which is much higher than SF-DORT,and has higher imaging efficiency while achieving similar high resolution imaging with SF-MUSIC.Furthermore,based on the study of ultra wide band frequency-frequency sample imaging theory,the SFF-DORT imaging method using joint space-frequency and frequency-frequency data is proposed.In the new method,echoes through single array measurement are used to build the new SFF-MDM.After performing SVD,left singular vectors for imaging are extracted by solving the fluctuation value of accumulated phase differences.Then,the signal subspace vectors are divided into multiple subvectors using two schemes where the target signatures are encoded by relative phase shifts between array elements and coarse frequency dependence,respectively.The final imaging function is constructed by accumulating subvectors corresponding to all antenna locations as well as coarse frequencies.Compared with the traditional FF-DORT method,the new method achieves better imaging resolution in both single-target and multi-target scenarios,and also has stronger anti-noise capability.The effectiveness of the proposed method is verified by numerical and preliminary experimental results.At last,with FDA as background,the approach of combining TR technique and FDA-MIMO radar to improve the system performance is explored.The FDA signal model and transmitting beampattern are derived,the beampattern properties and beampattern decoupling methods based on nolinear frequency coding are fully discussed.On this basis,the array model of FDA-MIMO radar is emphatically analysed,the signal reversal,retransmission and focusing in TR imaging theory are applied to FDA-MIMO radar received signal processing,and a new TR-FDA-MIMO joint range-angle estimation method is proposed.Simulation results show that the proposed method can effectively improve both the range and angle resolution of the output pseudo-spectral of conventional FDA-MIMO radar,and achieve more precise target parameters estimation.Besides,it shows greater robustness under conditions of different signal to noise ratio and antenna elements number.