Autonomous Ephemeris Prediction using Navigation Receivers
Published in , 1900
The prediction of ephemeris has been a long term problem being worked upon in the scientific community using the traditional methods of statistics and signal processing. In this paper, we present our machine learning-based approach for the autonomous prediction of ephemeris using navigation receivers. We have used the data provided by the Indian Space Research Organization (ISRO) for the duration of 30 days in the RINEX navigation file format and formalized the prediction of single day ephemerides using the previous 25 days’ navigation data (ephemerides). This approach includes handling inconsistent timelines in the data. We have used the ‘Forecasting at Scale - Prophet’ model for time series prediction. Using this method we devised a solution to reduce the Time to First Fix (TTFF) effectively. Read more
Download here