Dr.-Ing.

Kuhn, Thomas

Filters
Reset filters

Settings
Now showing 1 - 2 of 2
  • Publication
    Creating Trust in Collaborative Embedded Systems
    ( 2021)
    Akili, Samira
    ;
    Cioroaica, Emilia
    ;
    Kuhn, Thomas
    ;
    Schlingloff, Holger
    Effective collaboration of embedded systems relies strongly on the assumption that all components of the system and the system itself operate as expected. A level of trust is established based on that assumption. To verify and validate these assumptions, we propose a systematic procedure that starts at the design phase and spans the runtime of the systems. At design time, we propose system evaluation in pure virtual environments, allowing multiple system behaviors to be executed in a variety of scenarios. At runtime, we suggest performing predictive simulation to get insights into the system's decision making process. This enables trust to be created in the system part of a cooperation. When cooperation is performed in open, uncertain environments, the negotiation protocols between collaborative systems must be monitored at runtime. By engaging in various negotiation protocols, the participants assign roles, schedule tasks, and combine their world views to allow more resilient perception and planning. In this chapter, we describe two complementary monitoring approaches to address the decentralized nature of collaborative embedded systems.
  • Publication
    Towards Runtime Monitoring for Malicious Behaviors Detection in Smart Ecosystems
    ( 2019)
    Cioroaica, Emilia
    ;
    Giandomenico, Felicita Di
    ;
    Kuhn, Thomas
    ;
    Lonetti, Francesca
    ;
    Marchetti, Eda
    ;
    Jahic, Jasmin
    ;
    Schnicke, Frank
    Smart Ecosystem reflects in the control decisions of entities of different nature, especially of its software components. Particularly, the malicious behavior requires a more accurate attention. This paper discusses the challenges related to the evaluation of software smart agents and proposes a first solution leveraging the monitoring facilities for a) assuring conformity between the software agent and its digital twin in a real-time evaluation and b) validating decisions of the digital twins during runtime in a predictive simulation.