Studying on BDS positioning errors modeling by time series analysis

To improve the BDS positioning precision, the model of the BDS static positioning errors is presented by using time series analysis. The error series is verified to be non-stationary, so the non-stationary components, including the trend component and the periodic component, should firstly be extracted, and then we can build an AR (2) model to fit the residual stationary random component, parameters of the AR (2) model can be estimated by the least square method, Finally, the combination model of BDS positioning error sequence is constructed. The simulation experiment indicates that the autocorrelation of the modified positioning error reduces, and both the mean absolute errors and mean square errors of the reversed longitude sequence and latitude sequence decrease.