Integrated System Health Management (ISHM) technologies are mission-critical for deep space exploration. Space habitats are complex systems, made up of several subsystems such as Life Support, Communications, Data Acquisition, Thermal Control, Environmental Monitoring, and Electrical Power Systems. Deep space missions will face increased latency in communications with Earth. As such, it is critical that deep space aircrafts and habitats be capable of detecting anomalies early enough to mitigate the effects of communication delays. It would also be beneficial if the space crew had the tools to not just detect anomalies but also to help them resolve them without waiting for assistance from ground control. Traditionally, model-based reasoning techniques have been effective in monitoring the health of such operations, but the rising demand for rapid fault detection and response in deep-space habitats calls for autonomous monitoring software that is agile, scalable, and can respond to previously unseen events. Data-driven approaches using machine learning techniques can provide the ability to detect novel anomalies and to evolve over time. However, they are not as effective at generating explanations or tracing root causes as rule-based modeling systems. In this paper, we describe a case-based reasoning (CBR) approach to developing ISHM tool that combines the data-driven approach with a method for diagnosing and explaining them. CBR is an artificial intelligence (AI) technique that aims to solve problems by analogy. The system, Anomaly Detection via Topological feature Maps (ADTM), uses self-organizing maps (SOMs) as an unsupervised learning approach for modeling each individual case. ADTM's case-base models represent knowledge about different operational modes, including faults, as individual cases. The design of appropriate indexing mechanisms is crucial to the effectiveness of a case-based model. We describe two different approaches to case-base indexing and retrieval and compare their performance on a data set from a simulation of a CubeSat.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Case-Based Reasoning for System Anomaly Detection and Management


    Beteiligte:


    Erscheinungsdatum :

    2021-03-06


    Format / Umfang :

    1023556 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch




    11.0602 Anomaly Detection and Reasoning with Embedded Physical Model

    Institute of Electrical and Electronics Engineers | British Library Conference Proceedings | 2002


    ANOMALY DETECTION SYSTEM, ANOMALY DETECTION APPARATUS, AND ANOMALY DETECTION METHOD

    KOBAYASHI AYUMI | Europäisches Patentamt | 2023

    Freier Zugriff

    ANOMALY DETECTION SYSTEM, ANOMALY DETECTION METHOD, AND ANOMALY DETECTION PROGRAM

    MATSUDA TAKASHI / NISHIKAWA SHINYA | Europäisches Patentamt | 2023

    Freier Zugriff

    Sensor Tasking for Spacecraft Custody Maintenance and Anomaly Detection Using Evidential Reasoning

    Jaunzemis, Andris D. / Holzinger, Marcus J. / Luu, K. Kim | AIAA | 2018