Autonomous Navigation for a Lunar Satellite

Recent technological advancement and the commercialisation of the space sector have led to a significant surge in the development of space missions for deep-space exploration. In particular, lunar exploration has garnered heightened interest within the space community, driven by the prospect of substantial benefits and applications. This includes utilisation of lunar resources, lunar geology investigation and in-space technology demonstration. Currently, lunar missions mainly rely on Earth ground-based Guidance, Navigation and Control (GNC) operations involving human-in-the-loop processes. Although reliable, the ground-based navigation approach is prone to prolonged periods of communication delay, lacking real-time capabilities and autonomy. In addition, the booming growth of users in space will inevitably lead to saturation of ground slots, hindering the progression of space exploration. Reducing the dependence on ground operation by developing on-board autonomous navigation methods represents a potential solution for future lunar missions.  Achieving fully autonomous navigation for deep-space applications is a relatively new field of study. Lunar missions such as LUMIO and CubeX have been recently proposed and developed to demonstrate the feasibility of autonomous deep-space navigation based on optical images and X-ray pulsar measurements respectively. Although optical navigation represents the most mature technology for autonomous navigation, it typically results in low navigation accuracy when relying solely on optical measurements. In this work, autonomous navigation based on X-ray pulsar measurements will be investigated for a lunar-orbiting satellite. Three pulsars B0531+21, B1937+21 and B1821-24 are used as the navigation beacons and the satellite’s position can be derived from their signal timing measurements. An extended Kalman filter is used for state prediction and correction of the lunar satellite. A dynamic model is derived considering the main perturbations present in lunar orbits, including J2, J3, solar radiation pressure and third-body effects from the Earth and the Sun. Simulation results show that the position estimation can be achieved with a 3σ accuracy below 1km. The promising results demonstrate the capability of estimating the position of lunar satellites without relying on ground support. The significance of this study is anticipated to foster greater confidence in the implementation of X-ray pulsar based navigation, representing a foundational progression towards achieving fully autonomous navigation for lunar missions in the future.