直升机载ROSAR低空环境感知方法研究

发布时间：2018-02-26 16:49

本文关键词： 低空环境频率步进信号旋转式合成孔径雷达 Omega-K算法压缩感知　出处：《西安电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着直升机载等低空飞行器井喷式发展,低空飞行器的飞行安全及自主避险问题引起人们的广泛关注。针对低空环境感知中的感知能力受限于平台载荷、回波信噪比低、传统采样数据量大等问题,依次开展了步进频雷达低空障碍物超分辨一维成像、频率步进信号旋转式合成孔径雷达(ROSAR)成像、基于压缩感知的旋转式合成孔径雷达成像的方法研究。为了实现低空避险,提前感知危险信息,提高低空障碍物成像分辨率,提出了步进频雷达低空障碍物超分辨一维成像。首先,对原始回波信号去噪得到观测矢量,并构造字典矩阵;然后,通过改进的零空间调整算法获得一维超分辨距离像。该方法解决了常规压缩感知对噪声敏感、已知散射点个数及传统方法中IFFT后出现冗余问题,可用于步进频雷达低空障碍物检测与定位,是实现低空飞行器自主避险的前提。旋转式合成孔径雷达利用直升机旋翼的工作特点,可以实现360度环形成像,相比机头实孔径由于旋转合成孔径增大,能够检测更弱的目标和获得更高的分辨力。直接发射线性调频大带宽信号的ROSAR难以满足直升机平台对载荷尺寸和重量的要求。发射中心载频步进瞬时窄带信号的ROSAR,能够降低硬件复杂度,但带来了信号处理有效积累的难题。针对此难题,提出了脉冲间距离走动校正的频率步进ROSAR成像方法。首先,推导了频率步进信号旋转式合成孔径雷达的几何模型;然后,提出了频率步进雷达脉冲间距离走动及其频域校正方法;接着,采用Omega-K算法实现观测场景的高分辨成像;最后,对距离与方位分辨率性能影响进行详细的分析。计算机仿真结果验证了该算法的有效性以及高分辨性能。针对高分辨、高重频旋转式合成孔径雷达奈奎斯特采样数据量大与传统算法运算复杂度高的问题,提出了基于压缩感知的ROSAR成像方法。首先,在观测场景满足稀疏特性的前提下,方位向通过以稀疏孔径的方法获取降采样的原始数据;然后,在距离向使用匹配滤波方法实现距离向脉冲压缩;最后,在方位向则通过求解最小1l范数优化问题恢复方位向散射系数,从而完成ROSAR成像。与传统的距离-多普勒ROSAR成像方法相比,该方法在减少所需原始数据量的同时(可压缩至30%),降低算法复杂度,提升方位分辨率与压低方位旁瓣。计算机仿真结果验证了该算法的正确性和有效性。
[Abstract]:With the helicopter low altitude aircraft such as the development of blowout, low altitude aircraft flight safety and self hedging problems caused widespread concern. In low altitude environment perception perception is limited by the load of the platform, the echo signal-to-noise ratio is low, the problem of large amount of data sampling and traditional, in order to carry out the step frequency radar low altitude obstacle super resolution one dimensional imaging, step frequency signal of rotating synthetic aperture radar (ROSAR) imaging, the research method of rotary synthetic aperture radar imaging based on compressed sensing. In order to realize the low hedge, ahead of perceived danger information, improve the low obstacle resolution, put forward step frequency radar low altitude obstacle super-resolution imaging. The first one the original echo signal denoising, observation vector, and construct the dictionary matrix; then, obtain the one-dimensional super resolution range profile by adjusting algorithm. The improved null space The conventional method to solve the compressed sensing is sensitive to noise and redundancy problems after IFFT and the number of known scattering points in traditional method, can be used for step frequency radar low altitude obstacle detection and location, is the realization of low altitude aircraft self hedging premise. Synthetic aperture radar using characteristics of helicopter rotor rotating, can achieve 360 ring head imaging, compared to the real aperture due to the rotation of the synthetic aperture increases, can detect weak targets and obtain higher resolution. The direct emission of LFM signals with wide bandwidth ROSAR to meet the requirements of the helicopter platform on the load size and weight. The launch center carrier frequency step instantaneous narrowband signal ROSAR, can reduce the hardware complexity but the problem of signal processing of effective accumulation. In order to solve this problem, put forward the stepped frequency pulse range walk correction in ROSAR imaging method. First of all, is derived The geometric model of the stepped frequency signal of rotating synthetic aperture radar; then put forward the stepped frequency radar pulse and frequency range walk correction method; then Omega-K algorithm is used to realize high resolution imaging to observe the scene; finally, a detailed analysis of the effects of distance and azimuth resolution. The simulation results verify the effectiveness the algorithm and the performance of high resolution. For high resolution, high frequency rotary synthetic aperture radar data with the traditional Nyquist sampling algorithm with high computational complexity, the ROSAR imaging method based on compressed sensing. Firstly, based on the observed scene meet the sparsity, azimuth through access to sparse method the aperture to reduce the original sampling data; then, in the distance to use matching filtering method to realize the distance to the pulse compression; finally, in azimuth by solving the Small 1L norm optimization problem back azimuth scattering coefficient, thus completing the ROSAR imaging. Compared with the traditional range Doppler ROSAR imaging method, the method to reduce the required amount of the original data (at the same time can be compressed to 30%), reduce the complexity of the algorithm, improve azimuth resolution and low pressure range sidelobe. The simulation results verify the the correctness and effectiveness of the algorithm.

【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：V275.1;TN957.52

【共引文献】