Analysis of residual atmospheric delay in the low latitude regions using network-based GPS positioning

The atmosphere in low latitude regions is of particular interest to GPS researchers because the propagation of GPS signals becomes significantly delayed compared with other regions of the world. Hence this limits GPS positioning accuracy in equatorial regions. Although the atmospheric delay can be modelled, a residual component will still remain. Reducing, or mitigating the effect of residual atmospheric delay is of great interest, and remains a challenge, especially in equatorial regions. Analysis of relative positioning accuracy of GPS baselines has confirmed that the residual atmospheric delay is distance-dependent, even in low latitude areas. Residual ionospheric delay is the largest component in terms of both absolute magnitude and variability. However it can be largely eliminated by forming the ionosphere-free combination of measurements made on two frequencies. The residual tropospheric delay is smaller in magnitude but rather problematic due to strong spatio-temporal variations of its wet component. Introducing additional troposphere “scale factors” in the least squares estimation of relative position can reduce the effect of the residual. In a local GPS network, the distance-dependent errors can be spatially modelled by network-based positioning. The network-based technique generates a network “correction” for user positioning. The strategy is to partition this network correction into dispersive and non-dispersive components. The latter can be smoothed in order to enhance the ionosphere-free combination, and can be of benefit to ambiguity resolution. After this step, both the dispersive and non-dispersive correction components can be used in the final positioning step. Additional investigations are conducted for stochastic modelling of network-based positioning. Based on the least squares residuals, the variance-covariance estimation technique can be adapted to static network-based positioning. Moreover, a two-step procedure can be employed to deal with the temporal correlation in the measurements.