AI SUPERVISED PROPULSION SYSTEM FOR HIGH PRECISION FORMATION FLYING - EXPRO+

Objective: To design (including the algorithm definition), implement and test an AI learning machine, deriving optimised thruster input parameter commands, based on received signals from the AOCS/Payload sensors.Description: High precision formation flying is a design solution for some of the very demanding high precision observation missions like gravity observation missions, distributed synthesised antennas in LEO or space based multi-spacecraft telescopes. In such missions, the AOCS of the spacecraft requires a very high thrust precision in microN region or even lower. In the LEO Earth Observation Missions, the propulsion system encounters significant challenges such as providing a very high precision thrust in a very large dynamics and for a very long lifetime of the mission.The precision of the thruster response directly depends on the precision of the electrical values. A method to overcome mentioned issues is to calibrate the PPU and AOCS system in the flight, however, this is not optimal as it leads to observation discontinuity. To address this, the activity introduces a machine learning AI intelligence, which observes the response of the propulsion system tothe signals directly coming from the AOCS sensors like ultra fine accelerometers complemented by. e.g., the Star trackers or Fine Guidance Sensors (FGS). The AI then adapts the commends to propulsion system based on its previous responses and to react to the input signals from the AOCS sensors. AI can also be used for failure analysis and to implement measures to overcome the failure in EOP or science deep space missions. This would minimize the necessary time, link time and link sessions required. This activity encompassesthe following tasks: - Algorithm selection, where the suitable AI algorithm from different types like "Stochastic Gradient Descent", "Constraint Satisfaction" or "AI Supervised Learning Algorithms" will be chosen. - Implementation, where the chosen AI algorithmwill be implemented in an PPU microcontroller or FPGA simulator - Testing and development plan, where first the implemented AI algorithm in the PPU microcontroller or FPGA simulator will be tested on a thruster set-up and at the end a development plan towards high TRL levels based on the requirements from potential missions will be prepared.Procurement Policy: C(1) = Activity restricted to non-prime contractors (incl. SMEs). For additional information please go to:http://www.esa.int/About_Us/Business_with_ESA/Small_and_Medium_Sized_Enterprises/Opportunities_for_SMEs/Procurement_policy_on_fair_access_for_SMEs_-_the_C1-C4_Clauses