AI support for future mission operations: The case of BepiColombo on-board memory management

Abstract

The ESA-JAXA mission to Mercury, BepiColombo, is planned to be launched in the second half of 2018. The mission consists of two spacecraft, ESA’s Mercury Planetary Orbiter (MPO) and JAXA’s Mercury Magnetospheric Orbiter (MMO). One relevant aspect, in the context of the MPO, is theoptimization of the on-board science data storage and downlink, such that data-return-latency is minimized and SSMM overruns are avoided. To that end, an AI-based approach is currently under development to support both the definition of the SSMM and, in the future, science data downlink operations. The paper presents the tool and its algorithms and discusses the different aspects that have to be considered before deplyoing the tool in an operational environment, such as robustness of the solutions, different time granularity (from the current long-term plan to medium/short term plans), priorities, etc.