AI in the Sky: Proactive Interference Monitoring for the Ovzon Satellite Network (ASIMOV)

Purpose and goal

One of the main purposes of the project, to develop and evaluate an efficient machine learning algorithm in order to assess how well machine learning can be used in Overhorison´s spectrum monitoring, has been fulfilled. This is an important prerequisite for Overhorizon to develop and implement such an application. A successful exchange of knowledge in machine learning and satellite communication between RISE and Overhorizon was established through two workshops. All project goals have been achieved within the time frame of the project.

Expected results and effects

The project resulted in a machine learning algorithm that efficiently identifies interference in spectrum data, and in a successful exchange of knowledge in machine learning and satellite communication between the project partners. The project has lead to an increased knowledge of how well different machine learning methods work for this type of spectrum data and for Overhorizon a good prerequisite for implementing an application that monitors spectrum. The project has also resulted in the identification of new potential applications for AI in Overhorizon´s satellite service.

Planned approach and implementation

The project implementation has well corresponded to the project plan. The analysis and preparation of data was more extensive than anticipated, but also contributed to an important insight into how an application for interference identification can be developed. Knowledge was transferred partly through two workshops exploring the basics of machine learning and problem-specific methods and by RISE developing and sharing a program containing the steps needed to train a machine learning algorithm on Overhorizon´s data.