Abstract: Navigation satellite constellation design involves the selection of many variables for optimization, the purpose of which is to search a set of variables and generate a constellation to satisfy human requirements to the maximum. In this acticle, a scheme that took navigation performance and satellite cost as objectives is proposed to optimize a constellation. The configuration of the constellation is a hybrid composed of mediumearth orbit (MEO) and geostationary orbit (GEO) satellites. MEO satellites are used for global navigation, while GEO yields the navigation strengthening over China and the neighboring regions. The method of designing MEO and GEO is expounded. The approaches to calculate navigation performance and satellite cost are analyzed, and position dilution of precision (PDOP) is selected as the guideline of navigation performance. The concept of basic particle swarm algorithm and multi-objective optimization is introduced, and a modified multi-objective particle swarm algorithm (MOPSO) is proposed. The calculation process and test results of MOPSO are given. The mathematic model of multi-objective optimization of navigation constellation is discussed, and the definition domain of design variables is listed. By analyzing the optimization results of a navigation constellation, it can be shown that the analysis method and design scheme are feasible.

Key words: navigation, constellation, optimization, multi-objective, particle swarm algorithm