基于超短基线的GPS测姿算法研究

针对超短基线下姿态测量结果稳定性差及可靠性低的问题，提出一种更加可靠稳定的基于超短基线的全球定位系统（global positioning system，GPS）姿态测量算法。构建了扩展卡尔曼滤波（extended Kalman filtering，EKF）模型，并对扩展卡尔曼滤波模型中测量噪声矩阵采用基于卫星高度角模型，得出比较准确的浮点模糊度；联立了载波双差观测方程和伪距双差观测方程，并采用改进模糊度最小二乘去相关平差法（modified least-squares ambiguity decorrelation adjustment algorithm，M-LAMBDA）进行快速模糊度固定解算；利用了滤波法进行姿态解算。实验结果表明，与其他测姿算法相比，该算法可适用于超短基线，能够提高姿态结果的稳定性与可靠性，在基线为20 cm静态环境中，该算法模型航向角的误差方差为0.1左右，航向角的绝对误差为2°左右。在基线为20 cm动态环境中，该算法模型航向角的误差方差为2左右，姿态角的绝对误差为3°左右。

Research on GPS attitude determination based on ultra-short baseline

Concerning the problem of poor stability and low reliability of attitude measurement results under ultra-short baseline, a more reliable and stable attitude measurement algorithm of Global Positioning System (GPS) based on ultra-short baseline was proposed. Firstly, the extended Kalman filter (EKF) model was constructed, and the measurement noise matrix of EKF model was based on the satellite elevation angle model to obtain the more accurate floating-point ambiguity. Then, the carrier double difference observation equations and pseudo range double difference observation equations were combined, and the modified ambiguity least square decorrelation adjustment method (M-LAMBDA) was used for fast ambiguity fixed solution. Finally, the filtering method was used to calculate the attitude. The experimental results show that, the algorithm can be applied to the ultra-short baseline, and can improve the stability and reliability of the attitude results. In the static environment with a baseline of 20cm, the error variance of the heading angle of the algorithm model is about 0.1, and the absolute error of the heading angle is about 2°. In a dynamic environment with a baseline of 20cm, the error variance of the heading angle is about 2, and the absolute error of the attitude angle is about 3°.