Study On MIMO Radar Robust Waveform Design For Extended Target Detection

Radar extended target detection problem has drawn increasing attention from researchers.Since the mismatch of Target Impulse Response(TIR)may result in detection performance deterioration,this thesis considers the robust waveform design problem for extended targets with a colocated Multiple-Input Multiple-Output(MIMO)radar,aiming at improving the robust detection performance by transmitting high qualified waveform.Specifically,based on the Signal to Interference plus Noise Ratio(SINR)metric,robust waveform design problem is studied under different situations: 1)the radar is aided by TIR database;2)the radar is aided by imprecise TIR prior knowledge;3)the radar is aided by TIR with unknown phases;4)the radar is aided by a stochastic TIR model.Several algorithms are devised to design the robust waveform with Peak to Average power Ratio(PAR)constraint as well as Similarity and Spectral Compatibility(SSC)constraints.The performance of the proposed algorithms are analyzed theoretically and verified numerically.The main contributions of this thesis are summarized as follows:The annular TIR uncertainty set is proposed to describe the uncertainty of TIR.The current sets used to describe the uncertainty of TIR mainly include three categories: the discrete set,the spherical set and the stochastic set.Note that it is impossible to obtain the phase information when simulating TIR with electromagnetic methods,and the current sets are hardly to describe this situation exactly.This thesis proposes the annular TIR uncertainty model the TIR with unknown phases.Compared with the current uncertainty sets,the newly proposed annular TIR uncertainty set is able to match the situations where the phase of TIR is totally unknown and the scattering intensity of TIR is not exactly estimated.The annular TIR uncertainty set is used to assist the modeling of minimax waveform design problem with phase missed TIR.The maximizing robust detection probability criterion is proposed to design waveform for the stochastic TIR set.Maximizing the output SINR expectation is often used as the criterion nowadays when considering the waveform design problem for the stochastic TIR set.However,the output SINR expectation just show the first order statistical property instead of the distribution properties of the output SINR distribution.In the thesis,the probability density function of the output SINR is derived theoretically under Gaussian conditions,and maximizing robust detection probability criterion is proposed.The proposed criterion combines the output SINR expectation,variance and the desired SINR for radar target detection.Results show that,compared with the current criterion,the waveform designed based on the proposed criterion is able to adjust the output SINR distribution according to the radar operation environments,which is able to improve the robust detection performance.The two person zero sum game is constructed to explain the corresponding minimax waveform design models.The TIR dataset set,spherical TIR uncertainty set and annular TIR uncertainty set are used to model the situations where radar is aided by TIR database,imprecise TIR prior knowledge and TIR with unknown phases,respectively.In order to improve the robust detection performance of the radar,the waveforms are designed based on maximizing the worst-case output SINR over the TIR sets mentioned above.In addition,the two person zero sum game with MIMO radar and extended target being the two players is constructed to explain the corresponding minimax waveform design problems.The existence of Nash Equilibrium(NE)for the games is analyzed theoretically.Results show that the minimax waveforms over the spherical and annular TIR sets can be obtained indirectly by solving the NE strategies for the games.Algorithms are proposed to optimize minimax waveforms with SSC constraints.The current algorithms for minimax waveform design for extended targets mainly focus on energy or PAR constraint.However,the designed waveforms under energy or PAR constraint exhibit poor performance on pulse compression and spectral compatibility.Based on Minorization Maximization skills and dual theory,a series of algorithms to design minimax waveforms with SSC constraints is proposed.The convergences of the algorithms are proved and the computational complexities are derived theoretically.Results show that the waveforms designed by the proposed algorithms share excellent pulse compression and spectral compatibility properties,and achieve good detection performance.