Abstract:
Sampling of extraterrestrial bodies is an important development direction for future deep space exploration. Different from low Earth orbit missions, sampling of extraterrestrial bodies is usually restricted by many factors, such as long communication delay, limited energy and unstructured mission execution environment. On the one hand, the ground system cannot support such tasks in real time, which requires the spacecraft to have high autonomy. On the other hand, the limited computing resources of the system on the spacecraft and the high uncertainty of the mission environment lead to the complexity and low reliability of the fully autonomous design on the spacecraft, which still needs the decision support from the ground system to a large extent. The two contradictory aspects affect and hinder the smooth implementation of the task. According to this, in this article, through the analysis of celestial sampling system and its demand for artificial intelligence technology, a celestial sampling flexible system was designed and constructed based on artificial intelligence technology, to systematically support task execution task function modules by flexible configuration of the spacecraft and the ground system in the process, so as to gradually achieve high spacecraft autonomous ability to improve the efficiency of task execution. The system has been verified and applied in Chang’e-5 sampling package test, and the verification results show that the efficiency and reliability of sampling tasks can be significantly improved with the support of the system.