导航星座自主导航日益成为新一代卫星导航系统的主要研究方向。在系统地论述导航星座自主导航的信息处理流程的基础上，重点提出了导航星座自主导航的关键技术，包括卫星星历与时钟参数的长期预报技术、星间测距与通信链路的建立和维持技术、星座卫星自主时间同步技术、星座卫星自主星历更新技术、自主导航信息处理的鲁棒滤波技术、星座整体旋转建模技术，以及地球自转及极移参数的长期预报技术。详细分析了关键技术实施途径，论证了相关数学模型。最后，对星座卫星自主时间同步与星历更新算法进行了系统仿真，结果表明：通过星载滤波器处理星间双向测量数据，不断地修正卫星时钟和星历参数，能够高精度实现星座卫星时间同步和星历更新，初步展示星座自主导航信息处理流程及其关键技术算法的合理性和可行性。

If the autonomous navigation techniques are applied to a navigation constellation, the number of the ground control stations, the message injecting frequency from the ground stations to satellites and system maintenance cost may be reduced significantly. Meanwhile, the integrity of navigation messages can be monitored in the real-time mode and the survivability of the navigation system also enhanced. The autonomous navigation will be gradually a main research subject for the next generation navigation satellite system. Firstly, the information processing flow of the autonomous navigation constellation is described systematically in this paper. Secondly, the key techniques to realize the autonomous navigation for the navigation constellation, including the long-term prediction of the ephemerides and clock parameters, establishment and maintenance for the cross measurement and communication links, autonomous time synchronization, autonomous ephemeris updating, robust filtering, establishment of the constellation rotation models, long-term prediction of the earth rotation and polar drift parameters, are presented as an important part of the paper. Moreover, the approaches to actualize the key techniques are analyzed and the relative mathematic models are also demonstrated in detail. Finally, the autonomous time synchronization and ephemeris updating are simulated. It is clearly shown from the simulated results that by processing the inter-satellite measurement data with the on-board Kalman filters and updating continually the clock parameters and ephemerides, the time synchronization and ephemeris updating among the satellites can be implemented autonomously and highly-accurately. Therefore, the reasonability and feasibility of the information processing flow and key technical algorithms for the autonomous navigation constellation are validated preliminarily.