Description
The objective of the activity is to design, manufacture and test a power module that will be integrated in an autonomous spacecraftarchitecture.Targeted Improvements:Increase the level of onboard autonomy by 50% at overall platform level.Description: Current satellite operations require human intervention from ground, resulting in considerable inefficiencies and dependencies on communication window frequency, latency, and available resources, which become more prominent for large constellations. This activity aims to partially remove the current dependence of the spacecraft on ground stations by developing an intelligent power module for an autonomous satellite, based on Artificial Intelligence (AI) to autonomously process and interpret housekeeping and telemetry data and take decisions on needed actions in real-time. This is essential for management of future constellations. Current power subsystems can autonomously power up the system, however, for future autonomous satcom satellites, the FDIR (Failure Detection, Identification, and Recovery) system's capability to provide failure prognostics are crucial for the satellite autonomy especially in unforeseen circumstances. This is a very complicated technical issue that the traditional methods cannot cope with, thus, a new approach shall be developed targeting an autonomous system, based on AI, implementing health monitoring blocks in the power module, and interacting with the rest of the satellite, identifying in advance performance drift, preventing failures and keeping the power system operational in any flight circumstance. The concept is to train an AI system to identify combinations of housekeeping telemetry data series thatexhibit non nominal behaviours and take the right decision accordingly.The activity will develop an architecture for a power module for autonomous satellites. Primary functions to be autonomously implemented by the module will be trade-off taking resource requirements into consideration. Parameters, format and the commands to be generated will be identified, and the unit's management strategy will be developed.Finally, the algorithms needed to monitor the health and performance status, drifts and failure prognostics will be designed. A breadboard model of the power module will be designed, manufactured and experimentally tested to evaluate the developed concepts and algorithms and achieved autonomy.