CA3 :: Soft Computing and Autonomous Agents

UNINOVA has been increasing its experience, with past and on-going projects at ESOC, in radiation monitoring and prediction and the usage of this information in the assessment of Solar cells degradation (details available at http://www.gtd.es/fuzzy). This, along with our combined expertise in Artificial Intelligence, Data Mining, Machine learning, Fuzzy Logic and Decision Support Systems (mainly data warehousing), are behind this identified opportunity (see (Ribeiro, Pires et al. 2002) for an overview of these technologies and possible future applications).

Our experience of current practices of mission operations at ESOC is that operators can benefit from systems that deliver relevant information and knowledge, usually not integrated in the mission control environment. Space weather data is definitely included. In addition, space weather data is not used in standard documented mission control processes:

This perception is also confirmed by the assessment of current practices performed in the feasibility studies of (see (Horne 2000) – page 33). Operators cannot reliably identify space weather as a cause of problems since they do not have enough expertise in space physics and information about the spacecraft environment.
The identified user need is to supplement the flight control team with space environmental information - past, current and forecast - already translated into spacecraft operational impact assessment and, in combination with spacecraft telemetry data, into suggested preventive or corrective operational procedure(s) (as defined in the S/C Flight Operation Plan) for implementation.
With the opportunity implicit in the previous statements in mind, UNINOVA has envisioned a solution which consists of:
- Collecting historical and real-time space weather data coming from whatever sources are deemed relevant (opportunistic usage) for a specific operations context and integrate that data with the satellite telemetry and orbit.
- Transform the Space Weather raw data into information and knowledge at higher levels of abstraction.
- Develop a prototype system capable of delivering a set of space weather services designed for mission control purposes directly in the control room.

The user benefits include increased awareness of space weather cause & effect relationships; vis-à-vis on-board spacecraft health status; improved productivity and safety levels of satellite operations versus space weather environmental phenomena; increased science return and extended life time.

The SEIS prototype to be developed, will be used by the INTEGRAL and ENVISAT mission flight control operators (FCT teams) and integrates several key technologies:

Data Warehousing (decision support oriented database)

MOLAP Analysis (data exploration and correlation analysis reporting)

Artificial Neural Networks (general time-series forecasting)

Knowledge Based System (capture of domain experts’ knowledge)

While some of these technologies have already proven themselves on other domains of application (such as Data Warehousing and MOLAP Analysis – business domain), the latter (ANN and KBS) represent two blocks to be prototyped under the scope of the project.

Uninova would like to thank the follwing institutions for their support in
providing data, thus making the SEIS project possible:

UNINOVA wishes to thank Daniel Heynderickx (D.Heynderickx@oma.be) from
BIRA - Belgian Institute for Space Aeronomy (http://www.oma.be/BIRA-IASB/)
for his colaboration in the project, which allowed the use of BIRA's Space
Weather and S/C charging effects models contained in the SPENVIS package