An architecture concept for cognitive space communication networks

It is being increasingly recognized that the near-Earth space environment for civil, defense, and commercial sectors is rapidly becoming more congested and contested. Increasing bandwidth requirements coupled with increasing the number of satellites in LEO, MEO, and GEO orbits will lead to spectrum interference. Space communication networks will require spectrum sensing, dynamic spectrum allocation, and use of spectrum databases to mitigate these issues for the single link connectivity and cognitive networking techniques for the multiple link connectivity. Ground networks, which are capable of automatically connecting to various satellite networks, will need to be augmented with cognitive abilities as well. Emerging spectrum management approaches are becoming increasingly complex. In this paper, we propose an architectural approach based on the integration of technologies such as deep learning, cognitive radios, cognitive networking, and security. The approach enables a significant degree of automation in the space communication network. Several high-level aero and space scenarios where spectrum interference is going to be a key issue are identified. Details of proposed architecture will be systematically described from communications and security perspective. The current status of cognitive radio, networking, and machine learning applied to space communications will be summarized, and an approach to their integration and testing will be detailed.