To improve the generation of navigation products (precise orbits and clocks, ...) in terms of quality (accuracy and integrity), availability and robustness, based on the adoption of AI technology

GNSS observations, as taken by global GNSS sensor networks, are affected by a multitude of effects impacting the quality of the GNSS observations itself as well as the derived products. This is especially true for high accuracy products like precise orbits and clocks and also for corrections applied for real time applications. The effects impacting the GNSS observations range from unexpected GNSS behaviour, atmospheric disturbances (e.g. scintillation), environmental (e.g. multipath, antenna, ?) and seasonal effects up to those which are depending on the GNSS receiver behaviour. In this context, it has to be stressed that not all of the outlined effects can be modelled. This activity aims to develop algorithms exploiting AI techniques, which can be used to analyse GNSS observation data and derived products in real-time and post-processing. The AI algorithms shall allow the detection of events in the context of real time applications and systematic - long term effects in post processing. The outcome of the AI algorithm will be used for corrections of GNSS observations data, for deletion of GNSS observations, for the development of models in order to improve the GNSS observation quality and in addition, also the generated products (e.g. clocks, biases, real time corrections) and related quality can be monitored, corrected and flagged. Significant improvements in terms of observation and product quality are expected.This activity encompasses the following tasks:- State-of-art assessment- Design and development of algorithms exploiting AI techniques- Integration, testing and verification in overall GNSS observation algorithms- Conclusions and recommendations