传感网无源目标跟踪与传感器位置联合校正研究

Sensor network-based passive source tracking is essential for a variety of civil and military applications. In practice,the known sensor positions are almost inevitably subject to noises, which would significantly degrade the source tracking performance. Therefore, the investigation on the joint source tracking and sensor position calibration is of great importance. This proposal focuses on systems that utilize TDOA and FDOA measurements for passive source tracking. In particular, we shall investigate the following three aspects. First, we shall theoretically analyze the impact of sensor position errors on the source tracking performance. Its dependence on the positioning parameters,sensor array configuration and source motion trajectory will be established. We shall also show that ignoring the presence of sensor position errors would lead to severely degraded source tracking accuracy to motivate the study on the joint source tracking and sensor position calibration. Second,we shall develop novel joint source tracking and sensor position calibration methods that are in closed form, have low complexity and possess faster convergence to optimal performance than e.g., EKF. Third, we shall establish analytically the estimation error performance of the proposed algorithms and compare the results with the performance bounds to prove their statistical efficiency. We also plan to validate the theoretical developments via extensive computer simulations and hardware-in-the-loop simulations. This project is intended for laying the theoretical foundation for the applications of joint source tracking and sensor position calibration. It is expected that this project will also contribute to the development of Bayesian filtering and Kalman filtering theories.

传感器网络无源目标跟踪是许多民用和军用系统的基本要素。在实际应用中，传感器位置信息往往含有误差，导致无源目标跟踪性能恶化；因此研究无源目标跟踪和传感器位置联合校正对改善目标跟踪精度和传感器位置信息具有重要意义。本项目针对时差频差无源跟踪系统，对如下问题开展研究：（1）从理论上分析传感器位置误差对时差频差无源跟踪精度的影响，分析其与跟踪体制、传感器布局、目标运动轨迹和观测量精度等因素的关系，从理论上论证目标无源跟踪和传感器位置联合校正的必要性；（2）针对仅使用时差和同时使用时差频差的系统设计有闭式解、复杂度低、收敛快的无源目标跟踪和传感器位置联合校新方法；（3）从理论上证明新算法的统计有效性，通过计算机和半实物仿真验证理论研究的正确性。本项目属于应用基础研究，将为无源目标跟踪和传感器位置联合校正的实际工程运用提供理论指导和有效算法，并为贝叶斯和卡尔曼滤波等状态估计理论的发展提供新的思路。

基于目标信号到达时差（Time Difference of Arrival, TDOA）和到达频差（Frequency Difference of Arrival, FDOA）的无源定位跟踪是传感网的重要应用之一。前期研究表明：传感器位置误差可显著降低基于时差（TDOA）的无源定位精度。因此，研究无源目标定位跟踪和传感器位置联合校正对提升存在传感器位置误差时的系统性能具有重要意义。本项目考虑仅时差（TDOA Only）和时差频差（TDOA+FDOA）两种体制下的无源定位跟踪与传感器位置联合校正问题。.主要工作如下：.（1）推导了不同场景下目标定位跟踪的克莱默-劳下界（Cramer-Rao Lower Bound, CRLB），并通过数值仿真说明传感器位置误差、量测误差、定位几何构型等因素对定位精度的影响；.（2）针对声呐、卫星、地面站等不同平台，设计实现了多种TDOA only和TDOA+FDOA体制下具有闭式解、低复杂度的定位跟踪算法。理论分析证明：它们中的大多数是统计有效的目标位置或运动估计子。这些算法多数首先通过引入额外变量将非线性（Non-linear）的时差频差定位跟踪问题转为伪线性（Pseudo-linear）估计问题，然后使用估计误差校正等手段消去额外变量并提高定位跟踪精度。分析表明：上述算法的估计偏差（Bias）低于经典的两步定位法，能有效避免估计偏差的累积，适用于传感器位置误差条件下的无源定位跟踪。通过与课题组在传感器位置校正方面的工作相结合，获得了适用于无源目标跟踪和传感器位置联合校正的低复杂度算法，并应用于预研项目“对运动平台高轨卫星定位技术研究”中；实际数据测试表明：算法性能满足各项技术指标；.（3）开展了存在时统误差时的无源定位、TDOA+到达增益差（Gain Ratio of Arrival, GROA）体制下的定位、跟踪系统中移动传感器最优路径规划、压缩感知域的信号检测与重构、多目标无源跟踪等与项目密切相关的问题的研究，为项目成果后续应用打下了坚实的基础；.（4）在国际期刊（IEEE Signal Processing Letters、Digital Signal Processing、Ad Hoc Networks等）共发表SCI检索论文12篇；在《航空学报》、ICASSP等发表EI检索论文5篇；培养研究生6名，完成了项目预期任务。