Abstract: The traditional multi-epoch cumulative observational stochastic model usually ignores the time correlation of observations between epochs, which leads to its inaccurate characterization of the overall stochastic characteristics of observations. Considering the strong time correlation of GEO satellite observations in the BeiDou system, this paper proposes a method for constructing a time-dependent stochastic model suitable for BeiDou observations in multiple epochs. Based on the traditional multi-epoch stochastic model, the time correlation coefficient is directly introduced into the stochastic model, and the time correlation coefficient of each epoch observation is calculated by the residual after the difference between observation stations, and time-dependent stochastic model of BDS observation measurement under multi-epoch accumulation is generated, and its performance in integer ambiguity resolution is evaluated by actual experimental data. The experimental results show that the time-dependent stochastic model solves the problem of the overestimation of the lower limit of the PCF for the integer ambiguity in the traditional stochastic model to some extent, improves the Ratio value of integer ambiguity, and helps the integer ambiguity to pass the detection successfully. In addition, compared with the traditional stochastic model, the time-dependent stochastic model can effectively reduce the occurrence of missed detection and false alarm of integer ambiguity, and improves the reliability of integer ambiguity resolution.

Key words: time dependent, stochastic model, BeiDou, integer ambiguity, Ratio test