Study On Joint Estimation Of DOD-DOA And Polarization Parameters In MIMO Radar Based On Quaternion

The relevant research of Multiple-Input Multiple-Output (MIMO) Radar has beenthe hot research. Making up for the shortcomings of traditional radar system whichmakes MIMO Radar has a large advantage in areas such as target detection,identification and location. Radar target parameter estimation is a necessary part toachieve the targeting, tracking and imaging. The research of MIMO radar targetparameter estimation has an important significance.Bistatic MIMO Radar makes transmit and receive arrays split a large distance.Target localization has been achieved by transmitting independent distinguishablesignals to detect the targets and estimating the DOD/DOA. However, if there havemultiple close targets, the estimation performance will drop in MIMO Radar. Therefore,introducing the polarization sensitive array at the receive array to obtain polarizationinformation to distinguish different targets with different polarization, achieving thetarget parameter estimation with high resolution and the precise positioning objectives.As the extension of real and complex field, quaternion has been used in array signalprocessing widely. This paper introduces quaternion theory and establishes the signalmodel based on quaternion to improve the estimation performance. Since quaternionvector orthogonal brings a restriction relationship between the components of quaternion,the algorithm based on quaternion provides a more accurate estimation result and highertarget resolution. And the algorithm based on quaternion has a higher convergence speedthan the algorithm based on complex.This paper brings polarization sensitive array based on the estimation of thedirection-of-department, the direction-of-arrival in the traditional bistatic MIMO Radar,using quaternion and quaternion matrix theory to study the joint estimation of the DOD,DOA and polarization parameters in MIMO Radar. And taking into account the trendsthat MIMO radar turns into large MIMO radar, the estimation algorithm based on largerandom matrix theory in large MIMO radar has been studied.The work is supported by the project “Joint parameter estimation and targetlocalization based on polarization diversity in MIMO radar”(No.6107110) and the project “Robust target detection and estimation for MIMO radars based on largedimensional random matrix theory”(No.61371158), which are National Natural ScienceFoundation projects. The work in this paper is as follows:Firstly, based on the signal model in traditional bistatic MIMO radar, the complexsignal model has been established in bistatic MIMO radar using polarization sensitivelinear array and MUSIC algorithm, which is a traditional super-resolution algorithm, analgorithm has been proposed to joint estimate DOD/DOA and polarization parameters inMIMO radar, achieving the target estimation with high resolution.In the context of complex MUSIC algorithm, the joint estimation algorithm ofDOD/DOA and polarization parameter has been proposed based on quaternion. Thealgorithm uses quaternion and quaternion theory, and a quaternion signal model has beenbuilt in MIMO radar. The statistical covariance matrix of the received signal is achievedbased on quaternion and so does the QMUSIC spatial spectrum function. The algorithmimproves the estimation performance and the target resolution. The simulation resultsshow the effectiveness of the algorithm and QMUSIC algorithm shows the betterestimation performance comparing to the complex MUSIC algorithm.Considering the trends that MIMO radar turns into large MIMO radar, an algorithmof joint parameters estimation has been proposed based on large random matrix theory inlarge MIMO radar. On the basis of complex and quaternion model, G-MUSIC andG-QMUSIC spatial spectrum function has been constructed separately. While theelements of transmit and receive array are large and the number of snapshots cannotmeet the requirements, G-QMUSIC algorithm guarantee the effective target parameterestimation and comparing to the G-MUSIC algorithms, G-QMUSIC algorithm has ahigh resolution and better estimation performance. When the targets are close in air thatthe traditional algorithms cannot distinguish them well, G-QMUSIC algorithm can alsodistinguish the multi-targets precisely. And the simulation results prove that G-QMUSICshows the better estimation performance than G-MUSIC.