分布式易扩展的时频差集群定位理论与方法研究

The conventional centralized time-frequency passive location has been used in many applications. However, our studies found that the centralized system has the disadvantages of high complexity, poor scalability and weak viability. It cannot support the hot applications of unmanned aerial vehicle colony, space-based microsatellite, underwater distribution and other cluster passive location applications and it is difficult to meet the new reconnaissance development trend of high positioning accuracy, small support platform, fast response and strong survivability. As a result, this topic proposes the distributed and extensible time-frequency cluster location method, and studies the following three key theoretical problems: (1) For the optimization problem of distributed configuration, an optimization model based on Fisher information maximization is established, and a group matching scheme based on greedy rules is proposed to provide the theoretical basis for the site setting of distributed scheme. (2) For bias problem in existing TDOA/FDOA interpolation, an 2-D estimation model is constructed and an estimation method based on the combination of "soft" and "hard" constraints is proposed, which is aimed at reducing the bias of the parameter estimation and improving positioning accuracy in the following step. (3) For the problem of no common reference station in TDOA/FDOA based system, the target positioning solution planning function is established, and the radiation source localization algorithm using convex optimization and high-order compensation is proposed to realize the fast and efficient positioning of the radiation source. Finally, we will provide theoretical support for the cluster passive location through the research work.

传统集中式时频差无源定位已有较多应用，然而申请者研究发现，其复杂度高，扩展性差，生存能力弱，难以支持支持无人机蜂群、天基微小卫星、水下分布式等集群应用，难以适应定位精度高、支持平台小、反应速度快、生存能力强的新型侦察发展趋势。在此基础上，本课题提出分布式易扩展的时频差集群定位，研究以下三个关键理论问题：(1)针对分布式构型优化问题，建立基于Fisher信息量最大化的优化模型，提出一套基于贪婪规则的分组配对方案，为分布式方案的站点设置提供理论依据；(2)针对已有插值存在的有偏问题，建立二维分数倍估计模型，提出基于“软”、“硬”约束相结合的估计方法，减小参数估计的偏差，从而提高后续定位的精度；(3)针对时频差定位中无公共参考站的问题，建立目标定位解算规划函数，提出凸优化和高阶补偿的辐射源定位算法，实现辐射源的快速高效定位。通过课题研究工作，为时频差无源集群定位提供必要的理论支持。

无源定位系统是获取目标位置信息的重要手段。其中，时频差无源定位技术由于精度高受到广泛的应用。传统集中式无源定位存在系统要求高、扩展能力差、生存能力弱等劣势。为此，项目以多站无源定位系统为基础，研究分布式可扩展的无源协同定位技术，具体包括该定位系统中到达时间差（Time Difference of Arrival，TDOA）和到达多普勒频差（Frequency Difference of Arrival，FDOA）参数提取算法，以及基于上述定位参数的分布式平台目标定位算法。课题的主要工作和成果如下：.1.针对分布式时频差无源定位的构型设计，提出一种基于时频差参数估计与Fisher信息矩阵相结合的构型设计优化方法。2.针对传统参数估计算法受限于信号采样间隔和观测时间等不足，提出了一种高效的参数估计方法，借助广义延拓逼近法思想，利用标准的二阶锥规划模型进行求解。3.针对分布式平台，提出了一种基于极大似然的联合TDOA和FDOA定位算法，根据目标定位模型，利用MLE估计结果进行偏差补偿。4.针对站址误差条件下的定位，提出了一种考虑站址误差的分布式平台目标定位算法，引入站址误差，重新推导分析了MLE的理论偏差和定位算法的CRLB。5.针对现有高动态条件下定位参数估计算法性能损失严重、抗噪性弱、计算复杂度高的问题，提出了基于频域对称自相关和变尺度傅里叶变换的距离差和距离差变化率联合估计算法。.基于上述研究结果，课题组在Digtial Signal Processing、电子与信息学报等国内外重要期刊发表论文27篇，会议论文6篇，被SCI检索论文14篇，出版学术专著2部，申请并授权发明专利10项，协助培养博士生2名，硕士研究生1人。课题负责人在结题当年获得国家自然科学基金面上项目资助，获河南省自然科学优秀青年基金项目资助，获评副教授资格，并遴选为硕士研究生导师。