Low orbit regional enhanced navigation constellation for BDS3 design based on Bayesian optimization algorithm

Abstract

The application of Low Earth Orbit (LEO) satellite navigation can enhance geometric structure, increase observations and contribute to navigation and positioning. To improve the performance of the navigation constellation in China, this study proposes an optimized method of LEO-enhanced navigation constellation for BDS based on Bayesian optimization algorithm. In this paper, four different optimal LEO constellation configurations are designed, and their enhancements to BDS3 navigation performance are analyzed, including Geometric Dilution of Precision (GDOP), the numbers of visible satellites, and the rapid convergence of precision point positioning (PPP). Additionally, the enhancement advantages in China compared to other regions are further discussed. The results demonstrate that regional enhanced constellations with 70, 72, 80, and 81 satellites at an altitude of 1000 km can significantly improve the navigation performance of the navigation constellation. Globally, the addition of optimized LEO constellations has reduced the hybrid constellation GDOP by 19.0 %, 18.3 %, 19.9 %, and 20.3 %. Similar results can be obtained using the genetic algorithm (GA), but the computational efficiency of Bayesian optimization algorithm is 53.9 % higher than that of the genetic algorithm. The number of visible satellites of enhanced constellations in China has increased by more than four on average, which is better than that in other regions. In the PPP experiment, the convergence time of the stations in China and other regions is shortened by 83.0 % and 76.2 %, respectively, and the navigation performance of hybrid constellations in China is better.