【作者】 王怀军；

【导师】 粟毅；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2010， 博士

【摘要】 作为一种新兴雷达技术,多输入多输出(MIMO)雷达受到国内外科研人员的广泛关注。该雷达综合利用阵列与分集技术,能够引入远多于实际物理阵元数目的观测通道和自由度,并在信号层联合进行多通道回波数据处理,从而在目标检测、参数估计和成像识别等多方面性能上,相对于传统雷达得到了改善与提高。因此,MIMO雷达具有广阔的应用前景。论文针对MIMO雷达应用中的对空中(空间)目标的成像探测问题,开展了MIMO雷达成像算法研究,主要工作分为四个部分:MIMO雷达成像基础、MIMO雷达成像反向投影(BP)算法、MIMO雷达成像距离偏移(RM)算法和MIMO雷达成像旁瓣抑制与超分辨算法。MIMO雷达成像算法研究的前提条件是要弄清楚MIMO雷达成像的数据获取方式、天线布阵及阵列特性等基础性问题。因此,第一部分工作首先对MIMO雷达成像的数据获取方式进行了分析,进而讨论了MIMO雷达成像技术的天线布阵方式。然后,对比分析了MIMO雷达阵列、合成孔径阵列和实孔径阵列的空间采样能力,并讨论了它们的分辨性能,从理论上给出了MIMO雷达阵列的分辨率表示式。最后根据上述理论分析,构建了一套MIMO雷达成像原理性实验系统,利用实验数据进行了MIMO雷达成像性能评估。复杂的多收发阵列结构使得现有许多常用成像算法难以直接应用于MIMO雷达成像,这就需要探寻合适的MIMO雷达成像算法。第二部分工作首先将传统的BP算法推广应用于MIMO雷达成像,导出了MIMO雷达标准BP算法,它具有不受MIMO雷达阵列形式限制的优点。而后,基于时延曲线校正原理,提出了MIMO雷达TCC-BP算法,该算法大大降低了标准BP算法的运算量。综合传统的距离多普勒(RD)算法和BP算法,提出了MIMO雷达RD-BP算法,在保证成像质量的同时,RD-BP算法相比于标准BP算法和TCC-BP算法大大提高了成像处理的运算效率。最后通过MIMO雷达外场实测数据对上述所提算法的有效性进行了验证。第三部分工作则结合MIMO雷达阵列设计,在SAR RM算法的基础上,从空间谱域角度研究了MIMO雷达成像算法。首先,通过分析雷达成像与空间谱域填充的基本关系,提出了基于谱域填充的MIMO雷达SF-RM算法,并通过谱域支撑区分布分析对MIMO雷达的成像性能进行了评估。而后根据相位中心近似原理,进行了MIMO雷达天线阵列设计,进而提出了基于均匀等效线阵处理的MIMO雷达UELA-RM算法,该算法通过等效相位中心误差校正后可以方便地完成MIMO雷达成像。最后,结合收发正交线阵设计,提出了MIMO雷达OLA-RM算法,该算法能够有效实现窄带MIMO雷达二维“方位—方位”向成像。MIMO雷达成像BP算法和RM算法都是为了重建目标图像,它们不能够解决成像系统固有的高旁瓣和分辨率受限问题。以提高MIMO雷达的成像质量为目的,第四部分工作进一步研究了MIMO雷达成像旁瓣抑制和超分辨算法。首先基于空间频谱支撑区变形原理,提出了一种不损失分辨率且简单、有效的MIMO雷达成像旁瓣抑制算法—SRSR算法。在SRSR算法的基础上,分析了旁瓣抑制与频谱外推的内在关系,提出了MIMO雷达成像谱变形超分辨算法—Super-SRSR算法,与常规的超分辨率成像算法相比,Super-SRSR算法具有简单、高效、对噪声不敏感和非参数化的优点。最后,为了解决MIMO雷达空间频谱缺失情况下的超高旁瓣问题,提出了一种基于AR模型谱估计的MIMO雷达成像超高旁瓣抑制算法,该算法可以有效改善频谱缺失情况下成像结果的旁瓣性能。更多还原

【Abstract】 MIMO (Multi-Input Multi-Output) radar is an emerging technology that has drawn considerable attention. With judiciously designed antenna arrays and diversities, MIMO radar can obtain various channels and degrees of freedom, which are much more than the actual antennas. Through joint signal processing of the received data from multiple channels, MIMO radar has been shown to provide enhanced performance in detection, estimation, imaging and etc., so MIMO radar has significant potential for advancing the state-of-the-art of modern radar. Aimed at achieving the imaging of airborne (or spatial) targets, MIMO radar imaging algorithms are studied in this dissertation. The main contents fall into four sections, which are the elements of MIMO radar imaging, back projection (BP) algorithm for MIMO radar imaging, range migration (RM) algorithm for MIMO radar imaging, sidelobe suppression and super-resolution algorithms for MIMO radar imaging.The premise of research on MIMO radar imaging algorithm is to solve some basic problems of MIMO radar imaging, such as data acquisition, design of antenna array, array characteristic, and so on. Therefore, the data acquisition of MIMO radar imaging is analyzed in the first section. Next, the antenna array of MIMO radar imaging is discussed, and then the sampling abilities of MIMO radar array, synthetic aperture array and real aperture array are analyzed contrastively. Their resolution capabilities are discussed respectively and the resolution expression of MIMO radar array is demonstrated. At last, an experiment system for MIMO radar imaging is constructed. The excellent imaging performance of MIMO radar is validated through measured data processing and analysis.Many popular imaging algorithms are not applicable to MIMO radar because of its complicated multistatic architecture, so there is urgent need to explore suitable MIMO radar imaging algorithms. Based on conventional BP algorithm, MIMO radar standard BP algorithm is studied firstly in the second section. This algorithm is not restricted by the array architecture of MIMO radar. Through time-delay curve correction of MIMO radar range compressed data, a novel imaging algorithm called TCC-BP is proposed for MIMO radar. TCC-BP algorithm provides a significant reduction in computational burden. Combing conventional range Doppler (RD) algorithm and BP algorithm, RD-BP algorithm is proposed for MIMO radar imaging. Compared with the standard BP algorithm and TCC-BP algorithm, RD-BP algorithm dramatically improves the processing efficiency while maintaining the imaging quality. Finally, the effectiveness of the proposed imaging algorithms is demonstrated using the measured data by MIMO radar experiment system. MIMO radar spectral imaging algorithm is studied based on array design and RM algorithm in the third section. Firstly, the relationship between radar imaging and spectral filling is discussed, and MIMO radar SF-RM algorithm is proposed. Then MIMO radar imaging performance is evaluated according to the distribution of spectral support area. Based on phase center approximation and a new design of MIMO radar antenna array, UELA-RM algorithm is proposed for MIMO radar imaging, which can conveniently produce images after the correction of displaced phase center error. With orthogonal linear arrays, MIMO radar OLA-RM algorithm is proposed. This algorithm can provide two-dimensional image profile in both azimuth directions via a narrowband MIMO radar system.MIMO radar BP algorithm and RM algorithm are both for the image reconstruction. They can not overcome the inherent sidelobe artifacts and resolution limitation of MIMO radar imaging system. For the purpose of enhancing the quality of MIMO radar images, sidelobe reduction method and super-resolution imaging algorithm for MIMO radar are studied in the last section. Based on the reshaping of spatial spectrum, a novel algorithm called SRSR is proposed. SRSR can effectively reduce sidelobe artifacts without degrading the imaging resolution. The extrapolation of spatial spectrum due to sidelobe reduction is analyzed. Then a super-resolution algorithm in conjunction with SRSR is proposed for MIMO radar imaging. This super-resolution algorithm is called Super-SRSR, which has features of simple process, low computational load, non-sensibility to noise and nonparametric model. In order to solve the problem of huge sidelobes arising from gapped spatial spectrum of MIMO radar, a huge sidelobe reduction algorithm for MIMO radar is proposed. The proposed algorithm that uses autoregressive (AR) based spectral estimation technique is able to improve the quality of MIMO radar images.更多还原