JavaSTARS Drools Rule-Based Integrated System

During the fall of 2009 our senior design team for the Aerospace JavaSTARS project of 2009 – 2010 proposed a rule-based approach to aide the analysts using JavaSTARS software in detecting and identifying hardware or sensor faults onboard of a launch space vehicle.  Our vision was to use a rule-based expert system that integrates transparently with the JavaSTARS system.  In the span of the past academic year we have selected an open-source inference engine, designed the interface between this software product and the JavaSTARS system, implemented the integration between them, and developed and tested use-cases for the emerged integrated version of the JavaSTARSsystem.

Our developed system offers the functionality for the analysts to provide their expertise as logic statements that can automatically be evaluated against the appropriate data objects inside JavaSTARS. As the result of these evaluations, our system can detect deviations from the expected behavior at the rate of the incoming telemetry.

Our approach is data driven; it creates separate layers for the core JavaSTARS functionality and the expert system module.  This is accomplished via a set of rule files and spreadsheets that can be modified as needed without having to change the source code.  Additionally, our system can be extended to include a greater variety of input resources such as XML files, Web-based interfaces, and more.

Our system promotes increased expedience and accuracy at detecting anomalies with simple to use visual interface that alerts the analysts of possible malfunctions in a non-obtrusive way.  We believe that once our Rule-based JavaSTARS prototype had been promoted to a fully operational system, it would improve the critical mission assurance process in a real launch situation.

Student Team
  • Bahin Behdinian
  • Sanaz Bonakdar
  • Yuliya Yefimenko
Project Sponsor
Project Liaisons
Faculty Advisors