北斗三号中长基线RTK卡尔曼滤波算法

An algorithm of BeiDou-3 real-time kinematic Kalman filter on medium-distances baseline

针对BDS-3星座的中长基线实时动态定位,该文提出了一种基于组合观测值的实时动态(RTK)卡尔曼滤波定位算法。忽略差分模型后的大气延迟残差影响,以位置参数和组合观测值的双差模糊度作为状态向量进行滤波估计,并采用最小二乘模糊度降相关平差(LAMBDA)方法搜索固定模糊度,将模糊度固定后的组合观测值当作高精度的"伪距"观测值直接进行定位解算。通过山东省35~65 km基线数据测试结果表明:BDS-3超宽巷滤波定位的模糊度固定率基本接近100%,其平面精度为厘米级,高程精度为分米级;BDS-3宽巷滤波定位的模糊度固定率高达99.2%以上,其平面和高程精度均为厘米级。

Aiming at real-time dynamic positioning of the medium-long baseline of the BDS-3 constellation,an real-time kinematic(RTK)Kalman filter positioning algorithm based on combined observations was proposed.Ignoring the influence of atmospheric delay residuals after the difference model,positional parameter and the double difference ambiguity of the combined observations were used as the state vector for filter estimation,and the least-square ambiguity decorrelation adjustment(LAMBDA)method was used to search for the fixed ambiguity.Accurate"pseudorange"observations were used for positioning calculation directly.The 35~65 km baseline data test results in Shandong province showed that:the ambiguity fixation rate of BDS-3 extra wide lane filter positioning was basically close to 100%,its plane accuracy was centimeter level,and elevation accuracy was decimeter level;the fixed rate of positioning ambiguity was as high as 99.2%,and its plane and elevation accuracy were both centimeter level.