1. Using Coherence Inspectors to Determine the Reliability of Debriefed Epistemic Agents

     from Unmonitored Autonomous Team Missions  

       (International Symposium on Cognition Science and Engineering July 2013 Proceedings).

  1. Cameron Hughes, Ctest Laboratories

  2. Tracey Hughes, Ctest Laboratories

  3. Stephen Rhoden, PhD Youngstown State University


It is sometimes necessary to deploy  autonomous multi-agent teams  to perform analysis, interpretation,  threat or safety assessments in environments that are either too remote, hazardous, or physically impractical to send human beings. It may also be the case that  synchronous data communications  along with any real-time  video or audio  monitoring of the team may not be possible or at the very least not dependable.   In these situations if  we are able to debrief the team  at some point  how do we rely on any analysis, or assessment that the team makes if we  could not  monitor its  performance?   On teams that perform assessment or analysis  there is always the possibility that one or more members may make mis-identifications,  or false positives. Without the ability  to monitor , it may not be clear that  proper location  or substances  were analyzed and assessed. Without real-time monitoring and communications with the team it is also possible that  team concluded its analysis and assessment without  considering the entire environment. We could add to these problems  partial failure of any or all team members in terms of inter-team communication or subtask completion. In this paper we describe F.A.C.T  a  experimental heterogeneous multi-agent system consisting of  software agents and autonomous robots agents,  and  Coherence Inspector  a concept that consist of a   knowledge structure, and a knowledge engineering technique,  that  is used to determine the reliability of any analysis, interpretation or assessment that  F.A.C.T makes during the debriefing process.