Robustness of GNSS integer ambiguity resolution in the presence of atmospheric biases

Abstract

Both the underlying model strength and biases are two crucial factors for successful integer GNSS ambiguity resolution (AR) in real applications. In some cases, the biases can be adequately parameterized and an unbiased model can be formulated. However, such parameterization will, as trade-off, reduce the model strength as compared to the model in which the biases are ignored. The AR performance with the biased model may therefore be better than with the unbiased model, if the biases are sufficiently small. This would allow for faster AR using the biased model, after which the unbiased model can be used to estimate the remaining unknown parameters. We assess the bias-affected AR performance in the presence of tropospheric and ionospheric biases and compare it with the unbiased case. As a result, the maximum allowable biases are identified for different situations where CORS, static and kinematic baseline models are considered with different model settings. Depending on the size of the maximum allowable bias, a user may decide to use the biased model for AR or to use the unbiased model both for AR and estimating the other unknown parameters.