MIMO Radar Reconnaissance And Identification Based On Blind Source Separation

MIMO(Multiple input multiple output) radar is a new kind of radar system.Transmit/receive diversity or waveform diversity makes it possible to break the performance limit of traditional radar. Lower probability of intercept and higher performance of anti-jamming are achieved but also bring new challenges to radar countermeasure.This paper is mainly to study the reconnaissance and identification of MIMO radar as no radar countermeasure without reconnaissance. In this paper, firstly, the detailed working principle of MIMO radar is introduced from several key aspects such as the transmit/receive model, types of signal, matched filtering, virtual array and digital beam forming technology from which the characteristics of MIMO radar are concluded. Then the architecture and the working process of the reconnaissance system are studied. Then suitable receiver system, parameter measurements and sorting and identification method are analyzed based on the features of MIMO radar. The importance of intra-pulse feature analysis and extraction of intra-pulse modulation parameters is discussed. Then one method for modulation type identification of MIMO radar based on instantaneous auto-correlation is researched and simulation is presented.In order to describe MIMO radar specifically, more parameters are needed in the signal sorting and recognition, In view of the phenomenon that MIMO radar signals are mixed into one, this paper pays much attention to the whole process of blind source separation(BSS) algorithm and based on which a suitable model for radar reconnaissance is analyzed. Then an enhanced preprocessing step is presented which combines MDL with PCA to realize the estimation of the number of the original source and reduction of both dimension and noise, therefore the time of independent component analysis(ICA) will be shorter and results better. ICA algorithm based on real number was adopted in most former studies, but actually complex form which this paper prefers to is more common in real cases. At the end of this paper, three kinds of MIMO radar signal are simulated using Fast ICA algorithm, and the results show that BSS algorithm may be of greater importance to MIMO radar reconnaissance.