基于非线性检测前聚焦的高速微弱目标检测方法研究

Due to the rapid development of the high-speed weak targets, the security of nation defense is threated, and the detection performance of radar detection should be improved greatly. Based on the focus-before-detection technique, the detection performance can be improved with the same radar systems. The system complexity based on the linearity focus-before-detection is high because its computational cost has exponential relationship with the search dimensions, besides, the problem of blind speed side-lobe may appear, which will affect the rapid detection and engineering realization. These problems can be resolved at some cost of detection performances by the non-linearity focus-before-detection, which will be studied in this project. By introducing pseudo axis after transforming of the data, together with the utilization of the frequency convolution theorem, the robust and rapid non-linearity focus-before-detection algorithm will be designed for the high-speed stationary targets. For the high-speed and high maneuvering targets, the non-linearity focus-before-detects algorithms with high performance will be designed through constructing low order kernel function by introducing the non-uniform sampling and the auxiliary redundant data, and on this basis, the compensation of range migration and Doppler spread will be done jointly utilizing multi-dimensional data coupling elimination and multi-dimensional Fourier transform, which will eliminate the searching of velocity, the rapid detection for the high-speed and high maneuvering target can be done. Through the studies of this project, the performance of detection methods can be improved, and these methods could be realized in engineering, which has significance for the development of the detection methods for the high-speed weak targets.

高速微弱目标的快速发展对国防安全造成了重大威胁，对雷达检测性能提出了更高要求。检测前聚焦技术可在不改变雷达体制的前提下，改善雷达检测性能。线性检测前聚焦算法的计算量与搜索空间维数呈指数关系，使得系统复杂度太高，此外还存在盲速旁瓣问题，不利于目标快速检测和工程实现。非线性检测前聚焦算法可在牺牲一定检测性能的条件下有效解决上述问题，因此本项目将对其进行深入研究。针对高速平稳目标，本项目将转换数据域引入虚拟单元轴，并结合频域循环卷积定理设计稳健快速非线性检测前聚焦算法；针对高速高机动目标，引入非均匀采样思想和辅助冗余数据构建低阶核函数来设计高性能非线性检测前聚焦算法，并在此基础上，利用多维数据耦合消除技术和多维傅里叶变换联合补偿距离徙动和多普勒频移，达到消除速度搜索的目的，实现高速高机动目标的快速检测。本项目的研究可以提高目标检测性能，并有利于工程实现，对高速微弱目标检测方法的发展具有重要意义。

针对雷达高速微弱目标，非线性检测前聚焦算法可以在检测性能和检测效率上达到较好平衡，更有利于工程实现。本项目针对高速微弱目标非线性检测前聚焦算法展开研究，取得的重要成果包括：（1）针对高速平稳运动目标，根据雷达回波信号模型和回波数据在距离频域-慢时间域的分布，本项目提出了在距离频域设计核函数，并结合频域循环卷积定理快速完成信号能量的相参聚焦，实现了对高速平稳运动目标的稳健快速检测。（2）为了提高高速机动目标的运动参数估计性能，本项目首先提出了基于相参积累的三次自相关函数参数估计算法，该算法可以显著提高参数估计的抗噪声性能。为进一步提升计算效率，本项目在上述算法基础上又提出了一种基于相参积累的非均匀三次自相关函数参数估计算法，该算法通过引入非均匀采样思想构建核函数，可以在抗噪声性能保持不变的前提下，用更短时间完成参数估计。（3）针对高速机动目标检测算法中存在参数搜索过程使得计算量大的问题，本项目基于上面提出的三次自相关函数和非均匀三次自相关函数，提出了两种可以同时补偿距离徙动和多普勒扩散的检测算法，这两种算法可以消除参数搜索，实现高速机动目标的快速检测。本研究成果可为雷达目标检测领域提供一定的理论和工程借鉴。