复杂环境下多重信息约束的北斗农机连续导航模型研究

Satellite navigation and positioning technology play an important role in agricultural machinery navigation. However, the agricultural operation environment is complex and changeable. The quality of observation data is easily affected by surface topography, crop block and field reflection. This project is intended to solve the problem of high-precision discontinuous navigation of agricultural machinery, and to assist BeiDou navigation based on multiple constraints of agricultural machinery. The main contents are as follows: 1) Transform ionospheric delay from the oblique direction to the zenithal direction and build up an overall constraint, analyze the rules of multipath for BeiDou classified satellites and establish a fast cycle-slip detection and multipath compensation model based on classified satellites; 2) Excavate the trajectory and speed constraints information of agricultural machinery, introduce the inertial measurement aided information, thus establishing a loosely-coupled continuous navigation model constrained by multiple information; 3) Analyze the sensitivity of the posterior residuals and redundant observations to the outliers and their correlation, consider the variation of observation noise and establish an agricultural machinery continuous navigation quality control and pre-post feedback adaptive model. The implementation of this project has theoretical significance and broad application prospects for the improvement and development of the relevant theories of BeiDou and the advancement of "smart agriculture".

卫星导航定位技术在农业机械导航中发挥着重要作用，但农业作业环境复杂多变，观测数据质量易受地形地貌、作物遮挡和水田反射的影响。本项目拟针对农机卫星高精度导航不连续问题，基于农机多重约束信息辅助北斗导航进行研究。具体内容包括：1）归化斜路径电离层到天顶方向并整体约束，分析北斗不同类型卫星多路径变化规律，建立基于卫星分类的周跳快速捕捉和多路径削弱模型；2）挖掘农机轨迹、速度等约束信息，引入惯性测量辅助信息，建立多重信息约束的农机紧组合连续导航融合模型；3）分析验后残差和多余观测量对粗差敏感性及其相关性，顾及观测噪声变化特点，建立农机连续导航质量控制及前/后反馈自适应模型。项目的实施对我国北斗相关理论的完善和“智慧农业”的推进有重要理论研究意义和广阔应用前景。

项目针对复杂环境下农机不连续导航的问题，开展了复杂环境下周跳修复、磁暴环境下精密单点定位、卫星失锁环境下多源传感器融合、GNSS/INS紧组合及抗差EKF滤波等方面的相关研究。具体研究内容包括：（1）提出了基于周跳修复成功率的组合观测量系数优选模型，构建了强电离层环境下北斗三频周跳修复及其质量控制模型；（2）构建了基于大气延迟双约束的单频精密单点定位模型，阐明了磁暴环境对精密单点定位精度和收敛速度影响；（3）挖掘了大型农机轨迹、速度等约束信息，构建了基于转速/GNSS传感器融合的组合导航模型；（4）构建了GNSS/INS多源传感器紧组合融合模型，提出了改进抗差EKF滤波模型。研究结果表明：（1）整体上2EWL&1GF组合方式的周跳修复成功率较高，2EWL组合观测量的选取与周跳修复成功率不相关；对于低卫星高度角的IGSO和MEO数据，必须采用预报模型修正组合观测量以提高周跳修复成功率；即使在电离层活跃条件下，该模型周跳修复率仍接近于100%；（2）单频精密单点定位在静态和仿动态模式下，采用双约束模型在水平和垂直方向的收敛时间均最短；在磁暴复杂环境下采用对流层约束模型可大大提升收敛速度；（3）仅采用转速传感器数据滤波发散十分迅速，但在短时间内可获得厘米级导航定位解，可有效解决GNSS短时失锁引起的农机导航不连续问题；（4）遮挡复杂环境下GNSS/INS紧组合融合模型可保证导航连续性；改进抗差EKF滤波模型可减弱多种类型异常观测的影响，提高导航融合模型稳健性。项目的实施保证了复杂环境下农机导航的连续性，对“精准农业”推进有重要理论研究意义和广阔应用前景。