北斗三频观测值强电离层条件下的周跳探测与修复

采用伪距相位组合和无几何相位组合作为周跳探测量,分析了无几何相位组合的漏探率和误探率,并给出了该组合的适用条件。针对强电离层条件下误探率较高的缺点,提出了采用RBF(Radial basis function)神经网络模型对一阶电离层延迟变化量进行预测,并根据预测残差判断周跳是否发生,采用电离层延迟改正的相位组合周跳估值进行周跳修复。利用电离层活跃期间的北斗三频观测数据验证所提出的算法,实验结果表明:该模型即使在大磁暴发生期间也能够准确地探测出所有周跳并正确修复,不存在不敏感周跳组合,同时误探率在0.3%以下。

Cycle-slip detection and correction for BDS triple-frequency observations under strong ionospheric conditions

By using code-phase combination and geometry-free phase combination as cycle-slip measurement,the leakage rate and the mis-detection rate of the geometry-free phase combination are investigated,and the applicable conditions of ionosphere delay are given.In order to solve the terrible mis-detection problem under strong ionospheric conditions,a new methodology using RBF (Radial basis function) neural network is presented to predict the variation of the first-order ionosphere delay.The cycle slips are determined according to the prediction error and are corrected based on the computation by phase combination with the ionospheric delay taken into account.Experiments are made to verify the algorithm by using the BeiDou triple-frequency data under strong ionospheric conditions,and the results show that the proposed cycle-slip detection and correction method can accurately detect all cycle-slips even if with large geomagnetic storm,and can accurately correct all the cycle-slips.There are no insensitive cycle-slips,and the mis-detection rate can be reduced to ＜0.3％.