Multi-target Localization With MIMO Radar Via Tensor Decomposition

MIMO-based technology has received extensive attention with the development of communication technology and has been introduced into the field of radar research.The research of MIMO radar technology has gradually become an important branch of radar research.With the increasing radar system model,the structure of radar received signals is more and more complex.The traditional matrix-based radar signal processing method can not adapt to the new change.The received signals of MIMO radar has high-dimensional structure.The traditional matrix-based processing method does not make full use of the internal effective information of the signals.The tensor decomposition based processing method introduced in this paper can effectively utilize the multi-dimensional structure of complex received signals to avoid complicated operations,and can get better parameter estimation performance.Tensor decomposition is an emerging high-dimensional data processing method with excellent identification and high decomposition accuracy.We apply tensor decomposition to MIMO radar multi-target localization problem,which will greatly promote the development of relevant theories and the application of relevant technology.This paper studies the multi-target localization problem of MIMO radar based on tensor decomposition algorithm.The main contents are as follows:· We build a MIMO radar data model.In this paper,the bistatic multi-pulse MIMO radar data model and the single-transmitting base multi-receiving base MIMO radar data model based on three-dimensional coordinate system are constructed,and a multi-target localization method framework based on the above MIMO radar data model is established.The received signal model of the radar output is established,and the received signal is filtered to apply the basic tensor decomposition algorithm.We perform matrix reconstruction on the received signal to apply the traditional signal processing algorithm.The multi-target localization method for multistatic MIMO radar proposed in this paper does not require the assumption of mutual orthogonal waveform signals that traditional MIMO radars must have,which is of great significance for related research and application.· We apply tensor decomposition based and coupled tensor decomposition based algorithm to deal with multi-target localization problems of MIMO radar.We innovatively introduce Canonical Polyadic Decomposition(CPD)and Coupled Canonical Polyadic Decomposition(C-CPD)into a multistatic multi-pulse MIMO radar system,in which the received signals at each receiving array can be stacked to match the third-order tensor for CPD,and the set of these tensors share one matrix of factors and thus have a coupling relationship,which together follow the C-CPD model.We apply CPD and C-CPD based algorithms to these stacked tensors to estimate the correlation factor matrices,some of which can be used to extract location information of the target and to locate the spatial location of the target.The simulation results show that the established multistatic MIMO radar data model achieves good performance in the application of tensor decomposition based and coupled tensor decomposition based algorithms to deal with multi-target localization problems.