Information extraction in emergency management missions: an adaptive multi-agent approach

 With  increasing  demands  for  autonomous  agents  to  work  alongside  humans   in   emergency   management   response   (EMR),   considerations   of   translations  of  human  to  machine  language  (and  the  converse)  are  timely.  We  present   a   prototype   where   the   translation   is   dealt   with   by   restricting   communications to occur through a form of controlled natural language (CNL)  (Fuchs  and  Schwitter,  1995).  The  prototype  is  new  in  that  it  allows  for  communications  between  both  physical  and  virtual  autonomous  agents,  agents  are assigned different levels of autonomy, and it includes a level of information hiding  that  allows  for  information  to  be  passed  to  relevant  agents,  whilst  keeping  those  (humans)  involved  anonymous.  A  real-life  mission  is  then  used  to exemplify how information is retrieved and communicated in the prototype. Finally, some usability experimental results are presented.