Onboard FDIR based on AI techniques with data streams from Columbus and autonomous TM management control, Early technology project for future orbital missions

Artificial intelligence has the potential to become an integral part of space in the future to realize the autonomy of the system without ground control and crew interactions. The demonstrator developed in the scope of the ORBIT-STAR project supports the evaluation of scenarios and techniques for incorporating machine learning algorithms in an operational environment. The demonstrator provides an execution platform for hosting AI-based applications and establishes the interface with the COLUMBUS module of the International Space Station to receive sensor data and send event messages to ground and crew. The project was in the same time the pilot for the utilization of the platform for the evaluation of an anomaly detection and analysis use case for a selected COLUMBUS sensor data set. The applications were tested on the COLUMBUS Software Integration and Test Environment which delivered valuable information about the benefits and the pitfalls of the platform and the applications. The evaluation shows that if the platform can be further strengthened, it can be used to support experiments with AI applications and potentially even support operators in the maintenance of the station on-board.