<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-1">
<META content="MSHTML 6.00.6000.17023" name=GENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=#ffffff>
<DIV>Hi, </DIV>
<DIV> </DIV>
<DIV>We try to integrate a simulated user in the Let's Go dialog system. We have
written code and we</DIV>
<DIV>begin to see some interactions between the simulated user and the dialog
system. I explain you </DIV>
<DIV>that we have done : </DIV>
<DIV> </DIV>
<DIV> 1 - We developped a server called "AgentSimServ"
which is integrated as an Olympus agent in the dialog system. </DIV>
<DIV> We used the code of the "TTYRecognitionServer"
server for writing our server. </DIV>
<DIV> </DIV>
<DIV> 2 - The "AgentSimServ" receives frames from the
dialog manager, below an example of frame : </DIV>
<DIV> string frame = {c
launch_query :session_id "Default" :inframe "start<BR>
{<BR>
act request<BR>
object how_may_i_help_you<BR>
_repeat_counter 0<BR>
system_version 1<BR>
}<BR> end</DIV>
<DIV> </DIV>
<DIV> 3 - Then the AgentSimServ analyses the
frame and sends a sentence to the NLU server (Rosetta), via the hub,
like "go to the airport", or "yes" or "good
bye",...</DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV>I ask myself two questions : first, I would like to know if that we
have done is the good way to integrate a simulated user in the
dialog system. Secondly, I would like to know if it is possible to run the Let's
Go dialog system with only the dialog manager server and the simulated user
server, because we want to test the interaction between these two servers. If it
is </DIV>
<DIV>possible to run only these two servers, how we can do that ?</DIV>
<DIV> </DIV>
<DIV>Thank's in advance.</DIV>
<DIV> </DIV>
<DIV>Michel Ianotto.</DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV> </DIV>
<DIV>Michel Ianotto</DIV></BODY></HTML>