Information Technology Reference
In-Depth Information
As a result of these developments, realism of virtual soccer games has increased
very rapidly over the past years. Not only has the physical appearance of players and
environment become more and more appealing, also the technical movements as well
as the tactical decisions of the players have become much closer to reality. In contrast,
one particular aspect that has stayed a bit behind is the players' mental behavior, and
in particular their affective behavior. For instance, in state-of-the-art video games, the
only moments in which players' emotions are really apparent are right after a goal has
been scored. In such cases, players usually show some built-in emotional expressions
(e.g., cheering, or looking angry). However, complex emotional states that change
dynamically during the game and influence the players' behavior are usually very
limited. This is in conflict with the behavior of human soccer players, which has
shown extremely sensitive to the dynamics of their affective states [7]. Consequently,
existing soccer simulations are not as realistic (and appealing) as they could be.
To deal with this problem, the current paper presents a generic model for decision
making in relation to emotions and trust, and applies this to the domain of virtual
soccer. As also put forward in [16], endowing virtual soccer players with emotions
will enhance their believability, thereby making the game more engaging for
spectators. Both emotions and trust are mental states which have been found to have a
serious impact on people's decisions, in general as well as in sports context [7, 9].
The presented model takes the BDI framework [13] as point of departure, and extends
this with mechanisms to represent the dynamics of emotions and trust.
The remainder of this paper is structured as follows. The generic simulation model
is presented at a conceptual level in Section 2. In Section 3, this model is illustrated
for the context of soccer games, by showing a number of simulations generated in the
LEADSTO simulation environment. Next, Section 4 describes how the model was
incorporated into the virtual agents within the RoboCup 2D soccer environment, and
presents some preliminary results. Section 5 concludes the paper with a discussion.
2 Conceptual Model
In this section, the model for decision making with emotions and trust will be
described at an intuitive, conceptual level, using the agent-based modeling language
LEADSTO [2]. This language allows the modeler to integrate both logical
(qualitative) and numerical (quantitative) aspects. In LEADSTO, direct temporal
dependencies between two state properties in successive states are modeled by
executable dynamic properties. The format is defined as follows: let α and β be state
properties of the form 'conjunction of ground atoms or negations of ground atoms'. In
LEADSTO the notation
means:
If state property α holds for a certain time interval with duration g,
then after some delay (between e and f), state property
α
e, f, g, h
β
β will hold for a certain time interval of length h.
Here atomic state properties can have a qualitative, logical format, such as an
expression desire(d) , expressing that desire d occurs, or a quantitative, numerical
format such as an expression has_value(x, v) which expresses that variable x has value
v . For more details, see [2].
Search WWH ::




Custom Search