Section: New Results

Scenario Recognition

Participants : Annie Ressouche, Sabine Moisan, Jean-Paul Rigault, Ines Sarray, Daniel Gaffé.

Keywords: Synchronous Modeling, Model checking, Mealy machine, Cognitive systems.

For a long time, Stars strategy has been to favor the easy generation of activity recognition systems. These systems correspond to a succession of pattern matching and clustering algorithms, combined with adequate knowledge representation (e.g. scene topology, temporal constraints) at different abstraction levels (from raw signal to semantics). Due to the large range of application domains (surveillance, safety, health care, ...), we propose a generic approach to design activity recognition engines. Moreover, such domains require high dependability due to possible safety issues. Thus, our approach should also rely on formal methods to describe, analyze, verify, and generate effective recognition engines. We consider activity recognition engines as reactive systems that react to input events from their environment and produce output events in the form of alarms or notifications. Such engines are intrinsically real time, reactive and they evolve in discrete time. As a consequence, to recognize scenarios, we adapt the usual techniques of synchronous modeling approach to express scenario behaviors. This approach facilitates scenario validation and allows us to generate a recognizer for each scenario.

Our previous developments, on top of existing synchronous languages as Lustre and le (see secrion  7.22 ), were convenient for rapid prototyping. However, even if le is not a closed environment, it appeared as difficult as Lustre to customize, for efficiency reasons. This year, in the framework of Ines Sarray PhD thesis, we began to define a synchronous semantics for the future scenario language (see section  7.20 ). The idea is to generate automatically recognition engines at compilation time. The compilation itself is totally based on the semantics. To complete this approach we will rely on both our experiment with the le language last year and on the LE compilation process.