The primary role of this work package is to develop mitigation algorithms to tackle the effects of ionospheric disturbances, in particular the scintillation phenomenon, which frequently occurs in Brazil. These algorithms are tailored for the GNSS carrier phase based techniques, such as Real Time Kinematic (RTK) and Precise Point Positioning (PPP). After the in-depth review of the state of the art research, initial mitigation algorithms have been proposed and developed. The current work has concentrated on scintillation mitigation, using historical archived data from the CIGALA/CALIBRA monitoring network; the TEC related disturbances influencing GNSS positioning are also under research, using the TEC and TEC gradient modelling (WP200).
Starting from an analysis of correlation between the GNSS positioning performance and scintillation occurrence (WP100), several schemes were implemented and tested aiming to obtain a 'cleaner' observable. One approach is 'observation screening', in which the observations affected by scintillation are dynamically excluded. The idea of this method is to avoid using the observations highly affected by scintillation, and only use the 'cleaner' observations for positioning. Figure 7 shows a scintillation case study, where the amplitude scintillation index S4 is plotted against time. It can be seen that there is a clear effect of scintillation on this satellite, between ~04:00-05:00. The screening is applied in this time period, with a preset S4 threshold. In the preliminary results, a clear improvement can be found in the positioning performance. The second approach developed is 'observation weighting', which aims to improve the Least Squares stochastic model and maximize the usage of available observations. In this concept the observations are allocated different weights, depending on the level of scintillation to which they are subjected. A number of models, such as the Conker model (Conker et al., 2003) have been investigated to estimate the tracking error of the carrier phase and pseudorange observables, using the inputs of the monitored scintillation indices. These tracking errors are then used to estimate weights that are assigned to the corresponding observations (Aquino et al., 2009). The weighting approach can be considered as an improvement to the screening approach. Figure 8 illustrates the improvement of this approach in the positioning performance, where in this case, comparing to the standard solution, the 3D mean positioning error is reduced by 14.9%. It shall be noted that both screening and weighting approaches require the scintillation information of all observations, which can be obtained from the scintillation monitoring network and the output of WP200. This work package also actively works with the industry to implement the developed algorithms.
Figure 7: An example of a satellite observation affected by scintillation.
Figure 8: Positioning performance improvement with the weighting approach.