Research On Array Error Calibration For Multi-Channel Radar Systems

With the development of radar technologies, many multi-channel radar systems have been proposed in order to improve the detection performance: the distributed small satellites synthetic aperture radar (DSS-SAR) system and the multiple-input multiple-output (MIMO) radar system. By combining the signals from all the channels, the degrees of freedom can be increased, and more super performances can be obtained compared with mono-channel systems. However, the unavoided mismatchs among the channels can seriously degrade the performance of the combined signal processing, which have become the fundamental limitation for the multi-channel radar systems to achieve the theoretic performance in the applications. This dissertation mainly focuses on the study of array error calibration for the multi-channel radar systems: the DSS-SAR and MIMO radar systems. The research of this dissertation is summarized as follows:1. When the direction of the signal is unknown, the coupling of the Direction of Arrival (DOA) and the array gain-phase error becomes the main factor which causes the unavoid effects on the precision of the estimation. A new method based on the signals which are spatially and temporally disjoint with the separation angle known is presented to estimate the gain-phase error without knowing the absolute directions of the signals. The separation angle is used to separate the coupling of the DOA and the gain-phase error. The cost function is given to estimate both the DOA and the gain-phase error accurately. The estimates are acceptable without iterations and suitable for any array manifold.2. The gain, phase and position error estimation in the along-track configuration of the DSS-SAR system is researched. An array error estimation method based on the data in range time and azimuth Doppler domain is proposed first. In the method, it firstly estimates and compensates the gain error (which eliminates the interaction between the gain and position error estimations and guarantees that the position error estimate can converge to the true value), and then estimates the phase and position errors independently using the data of the zero Doppler bin (which avoids joint iteration between the phase error and the position error with low computational load). Since the signal-to-noise ratio (SNR) can be increased by range compression, a new method based on the date in compressed range time and azimuth Doppler domain is proposed. Its phase error estimation method can also be used in the case without Doppler ambiguities. And a novel gain error estimation method based on the DSS-SAR system is also introduced.3. To realize the function of achieving SAR image, GMTI and InSAR in the same time, both the along-track and cross-track baselines are needed which means that the system configuration should be three-dimensional. In such configuration, the along-track baseline is coupling with the cross-track one which makes it difficult to estimate them respectively. A new three-dimensional baseline error estimation method based on the data in SAR image domain is proposed. It can be used in the case when the along-track and cross-track baselines are both long without Doppler ambiguity existing and the very visible points are also needed. The along-track and cross-track baselines are sepatated by the image registration, and two kinds of baseline errors are estimated by image registration and subspace decomposition respectively. The influence of the estimation errors of the phase difference between different satellites on the estimates of the baseline errors is also dicussed.4. Compared with the conventional array, the gain-phase errors both of the transmitting and receiving arrays in the MIMO radar system should be considered. A new method that estimates the gain-phase error of the monostatic MIMO radar system is given. The“virtual array”similar to the conventional array can be obtained by using the orthogonal transmitted waveforms to match the received signals, and then the gain-phase errors both of the transmitting and the receiving arrays can be estimated. The remainder errors induced by the estimation course are also analyzed. The method is simple and feasible, and it is suitable for any array manifold.