Cognitive Autonomous Systems Laboratory
The WITAS Robotic Dialog Environment (RDE)
This section of the website is incomplete at present;
shaded items in the left-side menu are not yet available.
Major parts
The WITAS Robotic Dialog Environment (RDE) is a software
environment consisting
of programs and modules for the systematic development and
effective performance of dialogs between a human and a robot,
in particular using spoken or written natural language, often
in combination with still images and video. Each configuration
of the environment consists of three major parts: one or
more Autonomous Operator's Assistants, a Robotic World, and
a Software Infrastructure including in particular a
Video server.
Autonomous Operator Assistants
The dialog performing systems, as developed in RDE, are
being developed in the direction of Autonomous Operator
Assistants (AOA's).
(See the webnote under Robotic Dialog on the
CASL entry webpage for an explanation of this concept).
Several AOA's are being, or have been developed.
The main current AOA is called OPAS (for OPerator's ASsistant,
simply). It serves as a reference system, containing a
spectrum of facilities.
Specific research projects addressing a particular aspect
of the overall design of an AOA implement other systems
that pay particular attention to their particular topic of
study. They reuse parts and subsystems of OPAS so as to shorten
the development time.
The OPAS AOA consists of two main parts: the
Speech and Graphics User Interface (SGUI) and the
Dialog Performance System, called DOSAR. These are
described in separate webnotes that can be accessed from
the present left-side menu.
The Robotic World
Besides OPAS and other AOA's, the WITAS RDE contains an
interface to the actual WITAS helicopters, and to
several simulators that are used for simulating the
robot(s) when AOA:s or other dialog systems are being developed.
We use the term the Robotic World as a joint term
for these facilities. Amont the simulators, the Hazard
simulator is the most often used one. (Additional information
forthcoming).
Software Infrastructure
The software infrastructure contains software platforms for
the participating systems and for their communication,
and has the following parts:
- A development infrastructure, with
facilities for development and demonstration
support, for example, for recording and replaying dialogs.
- A communication framework defining the modes
and means of communication between subsystems.
- A video server that
provides for archive of video coming from the UAV, timestamping
of video frames, playback of selected video sequences, and
markup of the video frames with respect to location and
contents. The video server is an important resource for some
of the dialogs.
These three parts of the infrastructure are described under
separate menu items.
Additional, Experimental and Advanced Facilities
Besides the above, there are a number of other
programs and modules that play specific roles. Please refer
to the menu item 'Other Subsystems and Modules' for descriptions
of these.
The basic OPAS system provides a dialog for giving commands
to a UAV, querying it about its own state and the state of
the world that it observes, and specifying some simple
behavior policies. A number of additional services have been
implemented as extensions of the basic system, including a
simple planning/plan execution facility and a facility for
reflexive use of a logic of actions and change where the
system registers its own communication behavior in logical
terms. These extensions are still in a stage of development
and design iteration.
Posted on 2005-05-16 as part of the
CAISOR website.
[Version history].