Predicting Thermal Power Consumption of the Mars Express Satellite with Data Stream Mining

Orbiting Mars, the European Space Agency (ESA) operated spacecraft - Mars Express (MEX), provides extraordinary science data for the past 15 years. To continue the great contribution, MEX requires accurate power modeling, mainly to compensate for aging and battery degradation. The only unknown variable in the power budget is the power provided to the autonomous thermal subsystem, which in a challenging environment, keeps all equipment under its operating temperature. In this paper, we address the task of predicting the thermal power consumption (TPC) of MEX on all 33 thermal power lines, having available the stream of its telemetry data. Considering the problem definition, we face the task of multi-target regression, learning from data streams. To analyze such data streams, we use the incremental Structured Output Prediction tree (iSOUP-Tree) and the Adaptive Model Rules from High Speed Data Streams (AMRules) to model the power consumption. The evaluation aims to investigate the potential of the methods for learning from data streams for the task of predicting satellite power consumption and the influence of the time resolution of the measurements of thermal power consumption on the performance of the methods.

[1]  Saso Dzeroski,et al.  Tree-based methods for online multi-target regression , 2018, Journal of Intelligent Information Systems.

[2]  W. Hoeffding.  Probability Inequalities for sums of Bounded Random Variables , 1963 .

[3]  G. De’ath.  MULTIVARIATE REGRESSION TREES: A NEW TECHNIQUE FOR MODELING SPECIES–ENVIRONMENT RELATIONSHIPS , 2002 .

[4]  Nikola Simidjievski,et al.  Predicting Thermal Power Consumption of the Mars Express Satellite with Machine Learning , 2017, 2017 6th International Conference on Space Mission Challenges for Information Technology (SMC-IT).

[5]  Saso Dzeroski,et al.  Constraint Based Induction of Multi-objective Regression Trees , 2005, KDID.

[6]  Geoff Holmes,et al.  MOA: Massive Online Analysis , 2010, J. Mach. Learn. Res..

[7]  Saso Dzeroski,et al.  Incremental multi-target model trees for data streams , 2011, SAC.

[8]  Saso Dzeroski,et al.  Learning model trees from evolving data streams , 2010, Data Mining and Knowledge Discovery.

[9]  E. Walter,et al.  Multi-Output Suppport Vector Regression , 2003 .

[10]  Tapio Elomaa,et al.  Multi-target regression with rule ensembles , 2012, J. Mach. Learn. Res..

[11]  Saso Dzeroski,et al.  Predicting Structured Outputs k-Nearest Neighbours Method , 2011, Discovery Science.

[12]  Grigorios Tsoumakas,et al.  Dealing with Concept Drift and Class Imbalance in Multi-Label Stream Classification , 2011, IJCAI.

[13]  Zhi-Hua Zhou,et al.  A k-nearest neighbor based algorithm for multi-label classification , 2005, 2005 IEEE International Conference on Granular Computing.

[14]  Amanda Clare,et al.  Knowledge Discovery in Multi-label Phenotype Data , 2001, PKDD.

[15]  Reshma Khemchandani,et al.  Twin Support Vector Machines for Pattern Classification , 2007, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[16]  Rémi Gilleron,et al.  Learning Multi-label Alternating Decision Trees from Texts and Data , 2003, MLDM.

[17]  João Gama,et al.  Adaptive Model Rules From High-Speed Data Streams , 2014, BigMine.

[18]  Hai Zhao,et al.  Drift Detection for Multi-label Data Streams Based on Label Grouping and Entropy , 2014, 2014 IEEE International Conference on Data Mining Workshop.

[19]  João Gama,et al.  Adaptive Model Rules from Data Streams , 2013, ECML/PKDD.

[20]  Redouane Boumghar,et al.  Machine Learning for Spacecraft Operations Support - The Mars Express Power Challenge , 2017, 2017 6th International Conference on Space Mission Challenges for Information Technology (SMC-IT).