Workflows recognition through multi agents in surveillance systems

Abstract

Workflow management systems exactly enact business procedures and processes described in a process description language. This strict adherence to the prescribed workflow makes it impossible for the system to adapt to unforeseen circumstances. Surveillance systems have unpredicted information especially in difficult environments like the industrial ones. In this paper, we are presenting a workflow recognition architecture through the use of a multi agent system that controls and evaluates the recognized processes from the surveillance algorithms and adaptively creates environment warnings or alarms. The related methodology is based on Java technologies which are presented and latest innovations from the multi agents and workflow processes composition. © 2010 IFIP.