多误差源干扰下航空遥感位置姿态测量系统高精度实时估计方法研究

Position and orientation system (POS) is becoming a research hotspot for motion compensation of aerial remote sensing image. To realize high precision measurement of position and orientation under the complex environment with multiple disturbances, this project studies the POS device error mechanism combining with the characteristics of the carrier platform motion error, and establishes a higher-order nonlinear non-Gaussian stochastic error model. Based on the POS observability analysis, the optimum maneuver strategy suitable for aerial remote sensing task will be studied. By analyzing the coupling relationship of gravity disturbance and accelerometer measurement error and studying the distribution characteristics of gravity disturbance data, the accurate gravity disturbance model will be established with fast identifying the model parameters, and a high precision gravity disturbance method will be presented. By studying the mapping relation between the inertial sensor original data and the measurement information, an accurate prediction method of POS measurement information will be presented to solve the long GNSS outages problem. Then, analyzing the data fusion processing mechanism of POS, an integrated estimation structure based on the double filter is going to be designed to proofread motion errors effectively within the context of the synthetic aperture radar imaging quality. Through the research on characteristics of POS relative error initiated by periodical and pulsing filter-adjustment, an adaptive smoothing algorithm based on the dynamic node for filtering correction will be proposed to eliminate the influence of periodical and pulsing filter-adjustment on gliding property of motion compensation data. For the sake of verifying the validity and precision of these methods, a series of comprehensive experiments are carried out.

位置姿态测量系统（POS）已成为航空遥感实时成像运动补偿的研究热点，本项目研究POS在多误差源干扰下的器件误差特性机理，建立POS高阶非线性非高斯系统误差模型，分析系统状态可观测性，研究适合POS高阶系统误差模型的最佳载体机动策略。通过分析重力扰动误差与加速度计测量误差相互耦合关系，研究重力扰动数据分布特性演化机理，进行重力扰动精确建模和参数快速辨识方法研究，提出POS高精度重力补偿方法。通过研究POS原始惯性测量数据与量测信息之间的映射关系，提出卫星信号失锁下POS的高精度量测信息预测算法。研究POS数据融合处理机制，设计基于双滤波器的POS实时组合估计架构，从合成孔径雷达成像角度抑制运动误差效应，同时研究周期性脉冲式滤波校正引发的POS相对误差特性机理，提出基于动态节点的滤波校正值自适应平滑算法，消除周期性脉冲式校正对运动补偿数据的平滑性影响。最后进行综合实验验证上述方法的有效性。

针对位置姿态测量系统（POS）在航空遥感实时成像高精度运动补偿中的实际需求，本项目主要开展在多误差源干扰的复杂应用环境下POS的高精度实时滤波估计方法的研究工作。研究成果为航空遥感的位置姿态测量提供新理论和新方法，具有直接的工程应用价值。本项目完成的主要工作有：（1）针对传统惯性导航系统误差模型由于忽略惯性器件标定残差及简化误差源模型带来的不足，充分考虑刻度因子误差和安装误差的标定残差、近似计算误差等因素影响，建立了一个42维的高阶误差模型，实验结果表明，该模型可显著提高POS的滤波估计精度。（2）针对应用于机载SAR运动补偿中的POS数据处理方法的不足，提出一种基于滤波校正值平滑的POS数据处理方法，实验结果表明，该方法能够满足高分辨率机载SAR实时成像运动补偿的需求。（3）针对长时间卫星信号失锁情况下由于惯导误差累积而导致POS精度下降的问题，提出一种解决长时卫星信号失锁的量测信息混合预测方法，实验结果表明，该方法可对卫星信号失锁期间POS的滤波量测信息进行有效预测，从而辅助卡尔曼滤波器进行惯性/卫星数据融合处理。（4）针对POS导航计算中忽略重力扰动影响的问题，提出了一种精确重力扰动补偿方法，利用直接求差法获取一定精度的重力扰动值并建立一个合适的重力扰动模型，并基于此模型通过卡尔曼滤波器获得重力扰动的最优估计值，实验结果表明，该方法可有效提高POS的水平姿态精度。