Abstract |
The primary goal of an AmI environment is to help and support the people living in it; towards
that objective it should be able to identify a user need and act accordingly. Many research
approaches and commercial tools have focused on realizing this concept, which follows the
paradigm of the trigger-action model; however, the majority of them pose several limitations
(e.g., one trigger can be connected to a single action, artifact-oriented triggers mainly). The
domain of education would particularly benefit from an AmI environment able to monitor
students during their educational activities and intervene when deemed necessary to help,
support or motivate them so as to promote the learning process. Nevertheless, despite the
fact that the concept of the Intelligent Classroom has gained much attention from researchers
over the past decade, none of the approaches proposed so far offers a generic, scalable, fast
and easy way to connect triggers with actions.
Aiming to bridge this gap, this thesis presents a framework and an authoring tool that support
both developers and educators in defining the behaviors (triggers) that lead to context-aware
interventions (actions). Following an extensive literature review and an iterative elicitation
process – based on multiple collection methods such as brainstorming, focus groups,
observation and scenarios – the high-level functional and non-functional requirements that
both the framework and authoring tool should satisfy were identified. Based on those
findings, this work aims to equip the Intelligent Classroom with mechanisms that monitor the
learners' attention levels and intervene, when necessary, to (i) provide motivating activities
to distracted, unmotivated or tired individuals, or (ii) suggest to educators alternative
methodologies which would be beneficial for the entire classroom.
In more detail, the LECTOR framework offers a mechanism for identifying student behaviors
that require remedial actions and intervening when the students need help or support. This
mechanism relies on “if-then” rules -created either by developers or by educators - to define
the behavior of the classroom environment. In order to ensure scalability and simplify rules’
management, a three (3) step process for connecting a behavior with an intervention was
introduced. The first step requires the user to define a behavior, next the conditions under
which the behavior becomes a trigger have to be described, and during the last step
connections between a trigger and appropriate interventions are created. This decomposition
permits a behavior to be associated with multiple triggers, and a trigger with multiple
interventions that alternate depending on the context of use. Furthermore, in contrast to the
artifact-oriented recipes that are currently supported by the majority of such tools, LECTOR’s
rule structure supports the creation of user-oriented intervention scenarios.
Furthermore, LECTOR introduces a sophisticated authoring tool, named LECTORstudio, which
aims to support both developers and educators in creating rules that describe behaviors,
triggers, and interventions. However, since developers require further assistance in order to
integrate the appropriate building blocks necessary for programming the AmI environment
(e.g., actors, context, interventions), LECTORstudio also provides such functionality through
intuitive user interfaces.
Lastly, in order to further support the targeted end-users of the Intelligent Classroom
environment (i.e., students and educators), this work also features three (3) additional tools:
LECTORviewer, NotifEye and CognitOS. LECTORviewer and NotifEye aim to support educators
in having an overview of the students’ attention levels and providing their input regarding
ambiguous behaviors or scheduled interventions that aim to re-engage distracted, tired or
unmotivated students. CognitOS, on the other hand, is a sophisticated web-based working
environment for students that hosts a variety of educational applications, which constitute
the communication channels through which LECTOR presents the interventions
|