DrMUST: Automating the first steps of Anomaly Investigation


Published on

Once engineers have realised that an anomaly has happened, they face the problem of identifying the possible causes and other effects of this anomaly. DrMUST performs pattern matching to find similar behaviours across history (years) and correlation analysis to find which other parameters (out of 20 000) are involved in a given anomaly. DrMUST can be used not only for anomaly investigation but also to perform characterisations.
DrMUST is currently used by Venus Express and Planck both for pattern matching and correlation analysis. Flight Control Team reports 20% – 30% effort reduction in performing anomaly investigation (the other 70% - 80% effort is taken by taking corrective and preventive actions such as modifying procedures).

Published in: Technology
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

DrMUST: Automating the first steps of Anomaly Investigation

  1. 1. Dr. MUST: Doctor MUST Introducing innovative technologies in support of mission operationsThe Case Study – Once engineers haverealized that an anomaly has happenedthey face the problem of identifying whichare the possible causes and other effectsof this anomaly. In some cases, it is not Dr. MUST discover ASPERA as the cause of attitude errorspossible to know if this is the first Anomaly investigation is part of the routine allow to recognize the same word even ifoccurrence of an anomaly or if it diagnostic task of flight control engineers. spoken by different speakers. When an anomaly is detected (e.g. when a  Correlator allows to find the telemetryhappened before and went unnoticed. particular parameter crosses a noticeable parameters that are involved in a relevant threshold), it is common practice to search time period (e.g. anomaly) from a largeThe Solution – DrMUST allows to perform along the telemetry history for similar number of parameters (in the order of tenspattern matching to find similar behaviours behavioural patterns in order to characterize of thousands). In order to find whichand correlation analysis to find which other the anomaly. By analyzing the time periods parameters are correlated is makes to when the anomaly happened in the past, we assumptions: 1. parameters related to thesubsystems or parameters are connected may be able to identify its causes and anomaly behave similarly in all same-with a given anomaly. DrMUST can be eventually prevent it from happening again in anomaly periods; 2. Parameters related toused not only for anomaly investigation the future. The anomaly investigation process the anomaly will behave differently duringbut also to perform characterizations. can be very labour intensive: many different anomaly and nominal periods. The solving parameters need to be analysed to identify approach consists of scanning everyCurrent Status – DrMUST is used possible correlations with the observed parameter and suggesting to the users anomaly. those parameters with a similar behaviourcurrently by Venus Express and Planck during anomaly periods and differentboth for pattern matching and correlation Although DrMUST has been designed with behaviour during nominal periodsanalsys. The ESA Patent Group has the goal of supporting anomaly investigation; it can also be used to perform system or This is the Venus Express flight control teamdecided to protect DrMUST by filing a subsystem characterization. This process operational assessment:patent application in the European Patent helps engineers in identifying potential areas  Allows “googling” through spacecraft data:Office. of concern when operating the spacecraft in searching of similar occurrences or different modes. correlated occurrences.Operational Assessment – Flight Control  Impressive performance: queries run veryTeam reports 20% – 30% effort reduction Dr. MUST offers two main functionalities: quickly compared to manual searches.  Pattern Matching allows to find similar  DrMUST assists in tasks which are veryin performing anomaly investigation (the patterns in a given telemetry parameter labour intensive of data analysis. Before, itother 70% - 80% effort is taken by taking from a large time period (in the order of was required that the engineer wouldcorrective and preventive actions [e.g. years). When an anomaly is noticed, it is guess or hypothesize which parameters useful to understand if it is a really new could have a credible correlation to amodifying procedures]). anomaly or if it happened before (and specific behaviour and then would perform went unnoticed). The Pattern Matching the analysis to prove it or discard theProject Team – Developed for the ESA functionality of DrMUST allows to find correlation. DrMUST finds theseAdvanced Mission Concepts and when behaviours similar to a give one correlations for the engineer even theTechnologies Office by Black Hat S.L., happened in the history of a certain ones he had not thought about.Spain, and Solenix GmbH, Switzerland. parameter. Since behaviours are never exactly the same, the Pattern Matching Dr. MUST is a MUST Client in the sense thatESA/ESOC, Darmstadt, Germany functionality needs to allow for certain it uses time series data provided by theAdvanced Mission Concepts and Technologies Office flexibility to recognize similar behaviour. MUST repository. However, it can be easilyContact: Alessandro Donati The enabling technology is the usage of adapted to work with any kind of time seriesE-mail: Alessandro.Donati@esa.int speech recognition techniques as they data.Tel: +49 6151 90 2574