Study On Ultrawideband Thorugh-the-Wall Radar Imaging

Ultrawideband through-the-wall radar (TWR) is a kind of short-range surveillanceradar based on UWB technology. The capabilities of electromagnetic waves to penetratethrough nonmetallic materials provide a unique opportunity to detect, localize, image,track, and identify hidden objects behind the buildings or the barriers, with a wide rangeof applications in urban-warefare, anti-terrorism situations, hostage rescue situations,and other fields. In order to achieve good surveillance performace, imaging targets is animportant aspect. There are lots of algorithms proposed for through the wall radarimaging. Most of these algorithms are based on the back projection (BP) algorithm ordelay-and-sum (DAS) beamforming method, which can fast and simply image hiddentargets with a little prior knowledges. However, these methods with the wide mainlobeand high sidelobe level of the point spread functions (PSF) suffer from low imagingresolution and contrast. On the other hand, there are few articles on design of theantenna array and coded pulse modulation technology for improving through-the-wallimaging quality. This thesis thoroughly investigates the challenging problem of UWBthrough the wall imaging technology in the terms of imaging algorithms, MIMOantenna array theory and design, and pseudo chaotic pulse position modulationtechnology, which povide theoretical guidance for through-the-wall radar system designand application. The main content of this thesis can be summarized as follows:1. The principle of BP algorithm and DAS beamforming in through-the-wall radarimaging is analyzed, and the causes of problems of these algorithms are pointed. Twodifferent solutions to this problem based on the robust adaptive beamforming areaddressed. Firstly, an ultrawideband through-the-wall imaging algorithm that combinesthe robust Capon beamforming (RCB) with coherence factor (CF) weighting, namelyRCB_CF, is proposed. The received signals from all channels are alligned bycompensating for geometric attenutation, wall propagation attenutation and refractioneffects. Then RCB can improve the image quality in terms of resolution, and modifiedCF weighting can improve contrast and sidelobes by emphasizing the in-phase signalsand reducing the out-of-phase ones. Therefore, RCB_CF method results in simultaneousimprovement of imaging resolution and contrast, outperforming both BP algorithm andDAS beamformer. The other algorithm is an ultrawideband through-the-wall imagingmethod using cross-correlation weight of dual constrained robust beamforming (DRCB).Array is divided two subarrays alternately, and DRCB is applied to estimate the beamformings output signal each of two subarrays. The energy of their sum is regardedas pixel value in order to achieve higher resolution and much better interferencesuppression capabilities. Because these two beamformers give a well-correlatedmainlobe response and a different or uncorrelated sidelobe response, thecross-correlated coefficient of two beamformings output signals is weighted for eachpixel to suppress sidelobe so as to significantly improve constrast. The excellentperformance of the proposed methods is demonstrated by finite difference of timedomain (FDTD) numerical simulations and the experimentally measured data results.2. By observing the relation between the scalar factor of the minimum meansquared error (MMSE) beamforming weights and CF used for sidelobe suppression, andfinding the connection between RCB_CF method and MMSE beamfoming algorithm, ascaled MMSE beamforming is proposed. Due to the inhomogeneous wall medium,errors in antenna positions, quantization effects, etc, the robust estimation methods ofunknown parameters of a scaled MMSE beamforming are presented, and a generalformula of the scalar factor weighting a minimum variance distortionless response(MVDR) beamforming is given. The effectiveness of the proposed beamformer isdemonstrated in though-the-wall imaging by FDTD numerical simulations and theexperimentally measured data.3. The weight vector of thr eigenspace-based beamformer (ESB) is obtained byprojecting the weight vector of minimum-variance beamformer onto the signal subspaceof the measurement covariance matrix. The projection removes a residual noise-subspace componenet that considerably degrades the signal-to-noise (SNR) of thebeamformer output. Therefore, the ESB produces a SNR significantly higher than thatfrom minimum-variance beamformer, thus, it is a suitable method for noise suppressionin through-the-wall imaging. We analyze the performances of ESB, and point out theproblems of ESB adapted to through-the-wall imaging. Herein, ESB_RCB algorithm isproposed. The weight vector of the ESB_RCB is calculated by projecting the RCBweight vector with aiming at robustness against both finite-sample effects and steeringvector mismatches, onto a signal subspace constructed from the eigenstructure of theloaded sample covariance matrix. We show that the ESB_RCB method effectivelyreduces noise and enhances the contrast while the high resolution of RCB is retained.The effectiveness of the proposed method is validated in FDTD numerical simulationsand the experimentally measured data.4. The configuration of the multiple-input multiple-output (MIMO) array is a goodchoice of a sparse aperture array for UWB through-the-wall imaging system. MIMO array can achieve higher cross resolution and lower sidelobe level because it can getlarger array aperture and smaller array element spacing with the same number ofelements as the conventional single-input multiple-output (SIMO) array. According tothe imaging model, an equivalent array for UWB MIMO antenna arrays is analyzed andthe design strategies for UWB MIMO array are given. Simulation results demostrate theeffectiveness of the design strategies and the feasibility of their application inthrough-the-wall imaging system. In addition, UWB MIMO imaging algorithms usingcross-correlation coefficient weighting robust Capon beamforming is proposed. Theimaging results show that the proposed method can achieve high resolution andsignal-to-noise ratio in the image by FDTD numerical simulations and theexperimentally measured data.5. UWB impulse through-the-wall radar usually transmits pulse trains of verysmall width at regular intervals in time. However, it has range ambiguity and relativelypoor anti-narrowband-interference ability. The pseudo chaotic pulse-position-modulatedUWB radar system is proposed to solve these problems. An analytical expression of theambiguity function for the pseudo chaotic pulse-position-modulated ultrawidebandpulses is derived, and the range ambiguous characteristics are investigated. Thecapability of rejecting narrowband interference is also analyzed, which can provide atheoretical guidance for waveform design and performance analysis of through-the-wallradar system. A multipath scattering in an enclosed room comprising four walls ismodeled. The target detection capability, performace in narrowband interferenceenvironment, and imaging results of the pseudo chaotic pulse-position-modulatedtechnique are analyzed compared with the single pulse system and pulse trains system atregular intervals. In addition, when imaging targets inside a room, the interference ofmultipath will cause ghosts into the image, which can deteriorate the detection of target.We show that the aspect dependence feature of multipath ghosts is found inthrough-the-wall images. An image fusion method from multiple-view imaging isdeveloped for the suppression of ghosts in through-the-wall radar imaging. Validation ofthe proposed method with FDTD simulation data is provided.