Homepage | Links | Contacts

Duration: 9/2001-3/2003
Status: Concluded
Project Contact:
Rita Ribeiro

(Case 1 Executive Summary)
(Case 3 Executive Summary)

"Fuzzy Logic for Mission Control Processes" - Past Research Projects
Project Summary
ESOC/ESA contracted UNINOVA and GTD to demonstrate the advantages and/or limitations of soft-computing techniques (Fuzzy Logic, Neural Networks and Genetic Algorithms) applied to enhance ESOC/ESA mission control operations. The project is split into five cases, resulting in the implementation of Prototypes that will be validated operationally. UNINOVA is responsible for three of those five cases.

    Case 1: Envisat Gyros Health Monitoring System
    Development of a fuzzy expert system specifically designed for monitoring and diagnosing ENVISAT-1 satellite gyroscopes. Two main goals were defined for this system. One was to have a diagnostic tool capable of issuing intermediate warning alarms several days or weeks before a failure is detected by the satellite on-board software, in addition to the currently used Out of Limits (OOL) alarms. The second objective was to automate some of the analysis done by the Flight Control Team.

    The fault-detection techniques used for Gyroscope health are automatic checks performed by the on-board computer, out of limits alarms issued by the Mission Control System (SCOS 2000) based on spacecraft telemetry data, and off-line checks of gyroscope telemetry data performed by the Flight Control Team and Flight Dynamics Team. This project come out to cover the gaps that these techniques have, such as: are targeted for short-term detection and avoidance of false alarms and are defined as a set of crisp rules that either are satisfied or not.

    The project was designed to determine status of ENVISAT gyroscopes using full telemetry data and qualitative expert knowledge gathered during previous missions. Because Gyros rarely fail, being critical anyway, there is lack of historical data prohibiting predictive models application; moreover, there is no mathematical model to describe the gyroscope dynamics and the current decision fault detection is performed by humans. All these characteristics lead to the selection of Fuzzy Logic.

    Figure 1 illustrates the gyro typical frequency spectrum signature, which contains the Hunting Frequency Phenomenon (result of the physical characteristics of the gyro itself). Experts provided knowledge enabling the definition of a set of fuzzy rules to assess the gyro state. Below is an example of a rule.

    Figure 1 – Hunting frequency

    Figure 2 depicts the architecture supporting the application. Periodically the gyroscope telemetry data is injected into to the system incoming from S/C Performance Evaluation Tool (SPEVAL), which in turn is connected to the Mission Control System (SCOS2000). The data has to be pre-processed before being loaded into the database (the ODS operational data store) that will handle the data used by the application software. As is possible to see in the figure, there are two main users: (a) the S/C controller that uses the tool for operational verifications, and (b) the System Engineer to analyse more deeply the gyros health. Moreover, such tool is useful during pre-launch phase by providing the engineers with extra perception of the equipment.

    Figure 2 – Envisat Gyroscope Monitor Architecture

    Concluded and running in the control room, where is being used by the Envisat’s operators through the HMIs shown in Figure 3.

    Figure 3 – Envisat Gyroscope Monitor HMI

Case 3: Solar Array Performance Degradation
This case is concerned with monitoring the XMM solar array degradation and prediction of future degradation. Our goals are: (1) identify as many sources that cause degradation to the array as possible, (2) study each source effect independently, (3) exploring (in a broad sense) the possibility of a black-box predictive model for solar array performance degradation, and (4) development of a Decision Support System (DSS) to provide easy access and exploration capabilities of the data and the S/C component, space weather, etc.

Figure 4 depicts the developed architecture that is mainly composed of three main blocks:

  • Storage and AnalysisA Data Ware house supports all solar array telemetry data, and space weather that incomes from SPEVAL, LELA (application that provides the radiation level), and NOAA (a space environment service provider) site. The data is organized taking in account the XMM orbital position data. An OLAP server maintains the Multi-Dimensional cubes that are prepared to enhance the queries performed by the user.
  • Data MiningThis part is responsible for keeping predictive modes up-to-date, with the help of an expert, and to apply such models to data and so predicting the solar array degradation.
  • HMIAn OLAP Front-End provides navigation and search capabilities to the user.
  • Figure 4 – XMM Solar Array Performance Degradation Architecture
The predictive model was modelled using Recurrent Neural Networks (Elman Neural Network) usingthe telemetry (Max Delivered Power, Current) and environmental factors (e.g. Protons, X-rays). The results were predictions of max, min, and average value of the performance metric. The data is aggregated into blocks: half a day blocks by Max, Min, and average aggregations.

Figure 5 represents the actions to both build and apply the predictive model:

    Figure 5 – XMM Solar Array Performance Degradation Model build and use procedure
    Concluded and in delivery phase.

Case 5: Smart-1 Ion Thruster Diagnosis
The project will consist in performing diagnosis of the Smart-1 Ion Thruster.Detailed information will posted as soon as possible.

    User requirements definition

Research Areas
  • Fuzzy expert system
  • Neural Networks for Prediction
  • Data Warehousing / OLAP

F. Moura-Pires, R. A. Ribeiro, A. Pereira, F. J. Varas, G. Mantovani, A. Donati. Data quality fuzzy expert system. Proceedings of the 10th Mediterranean Conference on Control and Automation (MED02), Lisbon, Portugal, July (2002).

A. Pereira, F. Moura-Pires, R. A. Ribeiro, L. Correia, N. Viana, F. J. Varas, G. Mantovani, P. Bargellini, R. Perez-Bonilla, A. Donati . Fuzzy expert system for gyroscope fault detection. Proceedings of the 16th European Simulation Multiconference, Artificial Intelligence and Neural Network Track, 3-5 June, Darmstadt, Germany (2002).

Official web site: www.gtd.es/fuzzy

Apoio FCT – Fundação para a Ciência e a Tecnologia no âmbito da Unidade de Investigação CTS - Centro de Tecnologia e Sistemas, referência UID/EEA/00066/2013