Event Detection and Fusion Model for Semantic Interpretation of Monitored Scenarios within ASIMS Architecture - Foundations on Natural and Artificial Computation - page 522

Monitored Scenarios (ASIMS)”, to integrate all the abstraction levels of the in-
formation processing, from the sensor agent process and coherent fusion of agent
results to behaviour and situation identification. The architecture is currently in
the development and prototype stage.
This article first gives a general description of the architecture (section 2).
Then it describes the specific proposal for detection (section3),basedonfinite
automata models, and event fusion (section 4) in the intermediate abstraction
levels, and examples of use are given. The work finishes with the conclusions
and, in particular, with the approach for the future stages of the work.
2 ASIMS Architecture
Every automated interpretation system of monitored scenarios has a basic struc-
ture consisting of acquisition stages and information processing using abstraction
and decision-taking levels [1]. In our system, acquisition and part of the informa-
tion processing is done locally. The other information processing and decision-
taking is done centrally. In the system, the processing is distributed between
Remote Nodes and the Central Node. The Remote Nodes capture and process
the information in the environment around them. They send this information
to the Central Node, which is in charge of merging and processing it, and its
results are then passed to the decision-taking system (Figure 1).
Fig. 1. Image of sensors and Local and Global nodes structure
Each Local Node has a common storage system at all levels “Common Data
Model”. All the levels of the architecture (acquisition, segmentation, identifica-
tion, tracking, event classification and detection ) use this (Figure 2). The levels
are conceived as interchangeable modules within the architecture, which enable
modules to be easily changed at any moment.