Study On Methods Of Target Parameters Estimation For Bistatic MIMO Radar With Electromagnetic Vector Sensors

Bistatic MIMO(Multiple-Input Multiple-Output) radar, as an important branch of MIMO radar, placing the transmit array and the receive array separately, thus it obtains high diversity gain which improves radar performance. The transmitted signal and echo signal of bistatic MIMO radar propagate in space in the form of electromagnetic waves, and echo signals carry both the target angle information and polarization information of the target. The traditional algorithm only considers the DOD-DOA(Direction Of Departure-Direction Of Arrival) information in echo signals, without estimating and using the polarization information. Moreover, the long vector traditional algorithm is relatively complex and has large matrix dimensions, it also cannot guarantee orthogonality between the outputs of each component. Therefor, parameter estimation accuracy can be further improved by using the orthogonality of each component to describe the radar model on quaternion, and estimating target parameters in quaternion field.The following work has been carried out in this thesis.The thesis improves the two orthogonal electric dipole based bistatic MIMO radar model and derives the system model of electromagnetic vector MIMO bistatic radar. Comparing with the former model, echo signal in electric and magnetic fields can be sufficiently received, so that the amount of information is expanded. By using a separation coupling parameters idea and the model proposed in the paper, a RD-MUSIC(Reduced Dimension Multiple Signal Classification method) algorithm is given. Under the condition that parameters are automatically matched, DOD-DOA and polarization angle information are estimated step by step, which reduces the amount of computation, and enhances the practicality of the algorithm. For both bistatic MIMO radar models above, RD-MUSIC algorithm is used to estimate the parameters, the simulations show that the proposed model increases the RD-MUSIC algorithm parameter estimation performance by expanding the amount of information.In order to further reduce the workload of MUSIC algorithm with spectral peak searching, a rotating slice structure is designed using rotation invariant estimation(Estimating Signal Parameters via rotational Invariance Techniques, ESPRIT). Besides, an object angle information and polarization information joint estimation method is presented. Using ESPRIT algorithm, contrast experiments are done to compare parameters estimation performance of the two radars. The results also demonstrate that ESPRIT algorithm parameter estimation is more accurate, based on the proposed model.By using the quaternion theory, combined with the orthogonal structure of electromagnetic vector sensor, this thesis presents a quaternion transformation matrix and derives the quaternion model of bistatic MIMO radar with electromagnetic vector sensors. Based on the proposed model, this paper presents the quaternion field reduced dimension multiple signal classification algorithm(RD-QMUSIC). Simulations show that the united RMSE of the estimated parameters is smaller in quaternion domain than that in the complex domain, which means our algorithm is better.For the quaternion model of bistatic MIMO radar with electromagnetic vector sensors, in order to reduce computation in estimating DOD-DOA, this thesis proposes QESPRIT algorithm, which estimates DOD-DOA by the design of the rotating slice structure. On the basis of the knowledge of DOD-DOA, target polarization parameters could be estimated with QMUSIC algorithm. Simulations result show that the QESPRIT algorithm outperforms ESPRIT algorithm in the complex domain, in terms of estimation accuracy of DOD-DOA...