提出了一种适用于实时动态周跳修复的方法。该方法利用多普勒观测值(或伪距),包括产生候选值、筛选候选值、确定候选值3个步骤。与已有的方法相比,该方法无需对观测值进行预报,适用于连续周跳的探测。理论推导和试验结果表明,在发生连续周跳时该方法能够探测出所有的周跳。
In order to obtain consistent high-precision positioning results with GPS carrierphase measurements, cycle-slips must be correctly detected and removed or handled at the data processing stage. This paper proposes a new cycle-slip correction method. The approach includes generating and filtering cycle-slip candidates. It can be derived directly from the observation triple-difference time series under simple assumption. Compared with conventional methods, this method needs not prediction values of the measurements. The method is suitable for continuously cycle-slip. Theory and test show the method can correct all cycle-slip.