Research On Array Configuration And Parameter Estimation For Colocated MIMO Radar

Multiple-Input Multiple-Output(MIMO)radar can make full use of space diversity and waveform diversity to obtain more degrees of freedom so as to strengthen the target identifiability and enhance parameter estimation performance.The colocated MIMO radar,as one type of the MIMO radar,has small spacing between the array elements.It transmits orthogonal waveforms at the transmitter and applies the matched filters at the receiver to separate signals in each channel to form a large aperture.Thanks to the waveform diversity and the multi-channel signal processing,the degrees of the system freedom is greatly increased thus high parameter estimation performance can be obtained.This paper will study the array configuration and parameter estimation for colocated MIMO radar and the main work is as follows:(1)The parameter estimation of the target based on non-uniform linear array(NLA)configuration for MIMO radar is studied.The method of using the minimum redundancy linear array(MRLA)is proposed for NLA design of MIMO radar.Compared to the uniform linear array(ULA)with same number of antennas,the proposed method could increase the number of the virtual elements and enlarge the aperture of the virtual array thus enabling better parameter estimation performance and more targets to be identified.In order to reduce the computation load,a method for generating MRLA with large number of antennas is investigated.This method does not need the exhaustive search and has lower computation complexity.The MRLA obtained by this method also has low redundancy.(2)The parameter reconstruction based on compressive sensing(CS)for MIMO radar is investigated.Considering that the Gaussian random measurement matrix is hard to be implemented and is not satisfied for on-line optimization,a measurement matrix construction method for CS-MIMO radar is proposed based on chaotic random filter,which makes full use of the internal certainty and external randomness of the chaotic dynamical system.The coefficients of the random filter is designed by the chaotic sequence and the compressive measurement of the received signal is achieved.Compared to the Gaussian measurement matrix,the proposed method can realize on-line optimization for DOA reconstruction.Considering the loss of the randomness during constructing the two-dimensional measurement matrix with the one-dimensional chaotic sequence,an optimized measurement matrix design method applying the two-dimensional spatiotemporal chaos is proposed.It utilizes the spatiotemporal chaotic signal to generate the measurement matrix directly and can reduce the coherence of the sensing matrix thus increasing the accuracy of the DOA reconstruction.In order to obtain a further reduction of the coherence,a sensing matrix optimization algorithm based on singular value decomposition(SVD)is employed.This algorithm has quick convergence and can optimize the measurement matrix according to the information of the radar system task and target scene so as to reduce the estimation error and improve the scene recovery accuracy.(3)The compressed measurement and parameter reconstruction for CS-MIMO radar is investigated base on sparse random array.The method of applying the randomness of the element positions to realize compressive sensing is proposed for CS-MIMO radar.When the positions of the random elements follows one certain probability distribution,the kronecker product of the transmitting and the receiving array steer vectors can serve as the sensing matrix.The relations between the cross correlations of the sensing matrix,the Gram matrix and the array pattern are investigated in detail.Further more,when the random array is following the uniform distribution,the sensing matrix could satisfy the CS nonuniform recovery property is guaranteed.Compared to CS-MIMO radar with filled array,the proposed method can obtain more virtual array elements and achieve better recovery performance with less elements.Moreover,Since the proposed framework can not only avoid the additional random measurement matrix but also reduce the required elements,the complexity of the CS-MIMO radar system is greatly reduced.(4)The multi-parameter estimation for MIMO radar with circular array(CA)and L array is investigated.For the CA MIMO radar,the 2-D DOA estimation is achieved by employing the multiple signal classification(MUSIC)algorithm and the Capon method.The Cramer-Rao bound of the 2-D DOA estimation is derived specially.For the L array MIMO radar,the multistage wiener filtering(MWF)based Estimating Signal Parameters via Rotational Invariance Technology(ESPRIT)algorithm is given to achieve the joint estimation of the two-dimensional angle and Doppler frequency.Compared to the conventional singular value decomposition(SVD)based ESPRIT algorithm,the MWF based method has smaller computation load.