Study On Some Key Technologies Of MIMO Radar

As a new type of radar, multiple-input multiple-output (MIMO) radar consists ofmultiple transmitting antennas and multiple receiving antennas. The MIMO radartransmits orthogonal or partially correlated waveforms to obtain more degrees offreedom, as well as acquire angular diversity gain or to form the expected beampattern.Since MIMO radar has higher range and angular resolution, or diverse target viewangles, it can obtain higher SNR and combat the target radar cross section (RCS)scinsillation, which improves radar performances in target detection, multiple targetsidentifiability and parameter estimation. In addition, the antennas of MIMO radar withorthogonal waveform are omnidirectional, so it can keep all directions undersurveillance any moment and the expanding of the area under surveillance enableslonger coherent integrated interval. The performance of MIMO radar can be furtherimproved in many aspects through waveform design.This dissertation addresses some key technologies of MIMO Radar system. Theresearch is based on various configurations of MIMO radar array, such as MIMO radarwith widely separated antennas, the one with colocated antennas, the one with widelyseparated subarrays and beamspace MIMO radar. The research focus on the signalprocessing, target detection, target localization, velocity estimation and waveformdesign of MIMO radar, and it is summarized as follows.The detection methods of bi-static MIMO radar with colocated antennas thattransmit orthogonal waveforms are studied in several kinds of clutters. The Kelly, AMFand ACE detectors for bi-static MIMO radar system are developed. Theconstant-false-alarm-rate (CFAR) detectors in Gaussian clutter are derived underassumption of known clutter covariance matrix, unknown clutter covariance matrix anddifferent gain values between the primary data and the secondary data, respectively. TheCFAR detectors for bi-static MIMO radar in non-homogeneous clutter are also derivedunder assumption of known and unknown speckle covariance matrix in the case ofsufficient available secondary data and limited available secondary data. The detectionperformance of these methods are analyzed and compared under different cluttercircumstances via computer simulation. In addition, the compatible environment andfavorable condition of every detector are indicated.The detection method of airborne MIMO radar, which transmit orthogonalwaveforms with colocated antennas, in non-homogeneous clutter is studied. The modelof the clutter is constructed based on spherically invariant random vectors (SIRVs). Considering the fact that the clutter is non-homogeneous along the range dimension andthe clutter covariance matrix is singular, a modified asymptotical maximum likelihood(MAML) method is proposed to estimate the normalized covariance matrix of theclutter. A modified asymptotical generalized likelihood ratio test (MAGLRT) detector isdeveloped based on the MAML estimation of the clutter covariance matrix. Theinfluence of the clutter parameters, receiver noise and the number of transmittingantennas of the MIMO radar on the detector’s performance is analyzed and thesimulation results verify the superior performance of this method.The detection method for MIMO radar with separated antennas transmittingorthogonal waveforms is studied. According to the fact that, in a MIMO radar withwidely separated subarrays, a target locates in different range cells for different channels,a method which divides the detection area into identically shaped grids is proposed andtwo schemes to fix the test cell of the target in a grid for every channel are proposed aswell. The detailed approaches of both schemes are presented. Their performances ofboth schemes and the impact of the grid size on the performances are analyzed byadopting three detectors on radar with widely separated subarrays, the validity of thetwo schemes are proved.The accuracy of the target localization of MIMO radar as well as MISO radar isstudied. In the scenario that a target needs to be located by a small mobile platform, aradar system assisted by an array in the base station that operates alternately in MIMOand MISO mode is proposed. Geometrical dilution of precision (GDOP) of the targetrelative coordinate errors are formulated. The influence of the target location, thelocation of the small mobile platform and the number of the antennas in the array on theGDOP is analyzed under the two modes respectively.The Cramer-Rao bounds (CRBs) of joint estimation of the target location andvelocity for MIMO radar systems which combines widely separated and colocatedantennas is derived. Based on the CRBs, the impact of the number of subarrays in theMIMO radar and the target location on the estimations’ accuracy of the target locationand velocity are analyzed. It is proved that MIMO radar with more subarrays has betterand more stable performance for joint estimating the target location and velocity.The MIMO radar waveforms for target frequency responses (TFRs) estimation aredesigned. Statistic models for TFRs of multiple range-spread targets that are correlatedin slow-time are constructed. A waveform design method for MIMO radar is proposedbased on the steady state of Kalman filter. Simulation results are analyzed to illustratethat the waveforms optimized by this method can effectively reduce the errors of the estimations of the multiple range-spread TFRs correlated in slow-time.