With thousands of low Earth orbit (LEO) satellites to be launched in the near future, LEO mega-constellations are supposed to significantly change the positioning and navigation service for ground users. The goal of this contribution is to suggest and discuss the feasibility of possible procedures to generate the LEO orbital products at two accuracy levels to facilitate different positioning methods-i.e., Level A orbits with meter-level accuracy as LEO-specific broadcast ephemeris, and Level B orbits with an accuracy of centimeters as polynomial corrections based on Level A orbits. Real data of the LEO satellite GRACE FO-1 are used for analyzing the error budgets. For the Level A products, compared to the orbital user range errors (OUREs) of a few centimeters introduced by the ephemeris fitting, it was found that the orbital prediction errors play the dominant role in the total error budget-i.e., at around 0.1, 0.2 and 1 m for prediction intervals of 1, 2 and 6 h, respectively. For the Level B products, the predicted orbits within a short period of up to 60 s have an OURE of a few centimeters, while the polynomial fitting OUREs can be reduced by a few millimeters when increasing the polynomial degree from one to two.
Keywords: LEO; POD; mega-constellation; orbit prediction; positioning.