Agent-Based Development of Multisensory Monitoring Systems - Foundations on Natural and Artificial Computation - page 451

this case VigilAgent is used to control the access to a specific area. Assigning
human personal to carry out this task is a common solution. However, this
is an expensive solution where human's attention decreases after some elapsed
time working or watching video. Thus, it sounds interesting to perform this
task automatically via software which is responsible for collecting information
to inform when some anomalous situation occurs. In addition, it is relevant
to record normal situations to guarantee some usual activities; for instance,
that the current occupation capacity does not exceed above a threshold. So,
a company takes knowledge about economical losses arising for cunning users,
where it is necessary to have more personal to prevent risky situations. The
cost of the personnel is decreased because it is not necessary to have a guard in
every access control point. Notice that monitoring software does not replace the
security personnel but is complementary to offer a more ecient solution.
The rest of the paper is organized as follows. In section 2, an overview of
VigilAgent methodology phases is offered, justifying why Prometheus [5], INGE-
NIAS [7] and ICARO-T [3] technologies are integrated. Then, section 3 intro-
duces the case study used to demonstrate the applicability of VigilAgent . Finally,
section 4 offers some conclusions.
2 Description of the VigilAgent Methodology
The five phases of VigilAgent are briefly described next. (1) System specification
- the analyst identifies the system requirements and the environment of the
problem, which are obtained after several meetings arranged with the client;
(2) Architectural design - the system architect determines what kind of agents
the system has and how the interaction between them is; (3) Detailed design -
the agent designer and application designer collaborate to specify the internal
structure of each entity that makes up the system overall architecture produced
in the previous phase; (4) Implementation - the software developer generates and
completes the application code; and (5) Deployment - the deployment manager
deploys the application according to a specified deployment model.
At this point, it is worthy of notice several issues about this development
process. The first one is that phases named system specification and architec-
tural design in VigilAgent are the two first phases of Prometheus methodology
[5]. Another detail is that the third phase of VigilAgent (detailed design) uses
models of INGENIAS [7]. Finally, notice that code is generated and deployed for
ICARO-T framework [3]. Several reasons that are introduced in the following
have conducted to this integration.
Prometheus is significant because of the guidelines it offers to identify which
the agents and their interactions are. Another advantage of Prometheus is the
explicit use of the concept scenario which is closely related to the specific lan-
guage used in the monitoring domain. Indeed, a monitoring application is devel-
oped to deal with a collection of scenarios. Nevertheless, notice that Prometheus
last phase has not been integrated in VigilAgent because it focuses on belief-
desire-intention (BDI) agents and how the entities obtained during the design