北斗GBAS中的对流层异常完好性监测研究

With the popularization and application of dual-frequency ionospheric delay eliminating technology, the tropospheric anomaly has become the key factor that affects the integrity of BeiDou GBAS and seriously restricts the promotion and application of BeiDou GBAS in civil aviation. Aiming at this major need, this project will carry out research on the modeling, detection and error bounding for tropospheric anomaly in Beidou GBAS. First, based on the study on the spatio-temporal characteristics of multi-source tropospheric anomaly data, and the analysis of the statistical uncertainty of spatiotemporal tropospheric anomaly errors, a hierarchical correlation analysis and elimination method based on complementary error function is proposed to ensure the universality of the tropospheric model. Next, considering the characteristics of BeiDou mixed orbits, a multi-constellation tropospheric anomaly detection method based on orbital plane, an accurate setting method of omnidirectional threshold for the tropospheric anomaly detection statistics and a decoupling estimation method for simultaneous tropospheric anomalies in different directions are proposed to ensure the accuracy of the anomaly detection. Finally, an on-demand gradient ascent protection calculation method based on optimized risk allocation, a linear transformation and position domain bounding method of pseudorange domain tropospheric anomaly error based on the extreme-value model, and a heteroscedasticity modeling and bounding method of the thick-tail tropospheric anomaly error are proposed to ensure the integrity of the system. The models and algorithms proposed in this project can support the test and verification of the function and performance of the tropospheric anomaly monitoring in BeiDou GBAS. And the conclusions can be used as a reference for the draft of BeiDou GBAS international standards.

随着双频消除电离层延迟技术的普及应用，对流层异常成为影响北斗GBAS完好性的关键因素，严重制约其在民航领域的推广应用。瞄准这一重大需求，本项目开展北斗GBAS对流层异常的建模、检测和包络三个层面的研究。首先，研究多源对流层异常数据间的时空特性，分析时空相关对流层异常误差的统计不确定性，提出基于互补误差函数的分级相关分析和消除方法，保障模型的普适性；然后，利用北斗混合轨道特点，提出基于轨道面的多星座对流层异常检测方法、检测统计量的全向阈值精确设置方法、不同方向对流层异常同时发生的解耦估计方法，保障检测的准确性；最后，提出风险优化分配的按需梯度上升保护级计算方法、对流层异常伪距误差极值模型的线性变换与定位域包络方法、对流层异常厚尾误差异方差性建模与包络方法，保障系统的完好性。所提模型和算法可支撑北斗GBAS对流层异常监测功能和性能的测试和验证，所获结论可作为北斗GBAS国际标准草案的参考依据。

GBAS通过差分方法消除对流层延迟，但当出现异常天气时，该方法只能消除部分对流层延迟，造成残余对流层延迟误差，严重影响飞机定位精度和完好性。本项目从对流层异常影响可信评估、故障快速检测、风险优化分配、误差超紧包络四方面开展研究：（1）研究了非标称异常和管异常的时空特性及其对保护级的影响。结果表明非标称对流层误差的影响略大于管异常误差的影响。在距离为500m的模拟进近过程中，管异常引入的平均保护级增量为0.0823m，而非标称异常引入的为0.1597m，两者共同影响下的平均保护级增量为0.2157m，增长百分比为32.69%；（2）提出了基于轨道面的多星座对流层异常检测方法及全向阈值精确估计方法，建立了基于轨道面的低计算负荷故障检测模型。仿真结果表明，典型多星座配置下，应用本项目提出的模型可将计算时间算短66.6%；（3）提出风险优化分配的按需梯度上升保护级计算方法。实验结果证明：平均二分查找法和按需梯度上升法得到的HPL明显优于完全平均分配法。在满足总的风险要求下，平均二分查找法和按需梯度上升法的结果更符合理想值，优化可达8.6%；（4）提出了对流层异常厚尾误差异方差性建模与包络方法，利用极值理论对误差厚尾分布建模，提出了一种对厚尾误差极值分布的自主建模与包络方法，与传统的误差极值建模方法相比，该方法有效降低了模型参数的统计不确定性，使得误差分布模型更精确、误差包络更紧、保护级更优。