Cette Habilitation a Diriger des Recherches est le fruit des resultats, tant en recherche qu'en enseignement ou en transfert, de mon activite d'enseignant chercheur. Initiee au sein du Laboratoire d'Informatique, de Microelectronique et de Robotique de Montpellier (LIRMM, UMR CNRS/UM2) de 1991 a 1994, cette activite s'est ensuite concretisee au sein du Laboratoire de Genie Informatique et d'Ingenierie de Production (LGI2P) de l'Ecole des Mines d'Ales ou j'exerce depuis 1994. Le travail de recherche entrepris depuis le debut du Doctorat en 1991 releve de la thematique de la modelisation de systemes complexes puis de la verification et de la validation de ces modeles. Ceci a pour objectif d'assurer, ou a defaut de rassurer, le modeleur sur la qualite des modeles, sur leur pertinence vis-a-vis du systeme considere et sur le respect d'exigences qui ont preside a leur construction. La recherche a donc consiste au developpement d'approches de modelisation, de specification formelle de proprietes, de verification par preuve de proprietes au moyen de Graphes Conceptuels et de simulation comportementale. Les domaines d'application privilegies ont ete les systemes de controle commande repartis, puis plus largement la modelisation d'entreprise et tentent aujourd'hui d'integrer une dimension risque dans la modelisation d'entreprise et de s'ouvrir plus largement a l'ingenierie des systemes complexes. Les resultats sont des langages et un cadre de modelisation integre, un langage de specification baptise LUSP, une suite de mecanismes de preuve formelle et de simulation qui ont donne lieu a divers encadrements de theses, de travaux et a des transferts vers l'industrie. Enfin, l'activite d'enseignement a tente de rester coherente avec le profil de competence a la fois de producticien et d'ingenierie systeme acquis ou inspire par la thematique de recherche. Elle s'est deroulee dans le cadre de diverses Universites, Ecoles d'Ingenieurs ou de cursus specialises. Les resultats sont des propositions et l'accompagnement de thematiques nouvelles, une activite d'ingenierie pedagogique et une implication dans diverses responsabilites administratives.

https://tel.archives-ouvertes.fr/tel-00204981/document

Vérification et validation de modèles de systèmes complexes: application à la Modélisation d'Entreprise

The field of robotics is one of the most innovative in the last decade. Conventional industrial robots from the late 70 ́s are now only a tool on the production level. One of the oldest dreams of the robotic community – intelligent, mobile and humanoid robots – starts to become a reality because of the rapid development of “external” sensors. External sensors (e.g. visual, auditive, force-torque) offer intelligent robots the possibility to see, hear, speak, feel, smell like humans. Compared with conventional, unintelligent, industrial robots, intelligent robots can fulfill new, innovative tasks in new application areas.

Computer Aided Systems Theory - EUROCAST’99

Considering Model-Based Systems Engineering (MBSE) principles and needs, this article focuses on the design of Domain Specific Modeling Languages (DSMLs) aiming to link modeling, verification and validation (V&V) activities. The goal is to ease the work and increase the freedom and autonomy of experts from various domains in the process of creating and managing system models then to supply experts involved in analyses and decision making processes with models characterized by highest level of confidence. This paper introduces and illustrates a tool-equipped approach, named “xviCore”, that provides MBSE experts with natively verifiable, executable and interoperable DSML named “xviDSML”, i.e., DSML that can be formally and directly checked and simulated requesting no huge efforts or skills.

Towards V&V suitable Domain Specific Modeling Languages for MBSE: A tooled approach

The Interpreted Sequential Machine (ISM) model handles a new approach for modeling and formal verification of discrete complex systems.

Formal Representation and Proof of the Interpreted Sequential Machine Model

https://hal.archives-ouvertes.fr/hal-01930280/document

MBSE and V&V: a tool-equipped method for combining various V&V strategies

The complexity of systems requires that validation methods are set in the designing phase as well as during their use or maintenance. We propose a modelling approach of the system which allows us to bring into play some methods for the validation of temporal and functional properties. The model that we have defined, called Interpreted Sequential Machine, is based on the concept of Sequential Machine and avoids the main limitation (combinatorial explosion of the number of states when introducing any new data) by separating the purely sequential part of the system from the data and the operations on the data. The validation of the complex system thus modelled consists in:


expressing the behaviour of the system by a set of symbolic, logicotemporal formulAE,


carrying out automatic proof procedures on these formulAE.

Temporal and Functional Verification of a Symbolic Representation of Complex Systems

In order to support the design of complex systems such as optronic systems, and especially the function-processing oriented part of this work, we propose a method based on a connection of stages. This is aimed at making the representation of the system and its properties both formal and checkable. These formalizations can be used to manage the real verification of the system and determine the agreement between its model and the expected properties. The principle is to follow the first steps (corresponding to the design phase) of the traditional V-cycle of system development, and to enrich each step with formal features: the first step is the elaboration, based on initial requirements, of a function processing oriented specification. The first result is then a formalized (functional and non-functional) requirements set. It allows the system designer to construct, using the generic model and properties inventory, the precise model of the system (i.e. an instantiated model, called MOTI), and the list of properties to be verified. This MOTI is then (automatically) translated into a so-called MOTIF expressed in a formalism dedicated to formal proof (required for formal verification and validation processes). Concurrently, the expected properties (which have been extracted from the generic properties inventory enriched by specific characteristics such as user-defined values or application-specific parameters), are translated into the accurate formalism for automated proof.

A process for improving multi-technology system high level design: modeling, verification and validation of complex optronic systems

The task of anticipating the behavior of a system requires to have a model as a basis for reasoning This statement raises several questions and problems, especially concerning the knowledge the user has got of the system, and his ability to describe the aspects of the system he wants to study This point also involves that some models are available, and consequently, that some associated methods for predicting the future evolution exist Furthermore, the description or modeling is sometimes reduced to an observation of the reactions of outputs when inputs are applied, or can be more detailed Moreover, there exists a large range of ways to handle and represent the “time” factor The choice can be determining for the possible reasoning processes This paper discusses these aspects and presents a particular solution for systems described from a discrete time point of view

Temporal logics for analyzing the behavior of systems

Verification and validation appear to be crucial tasks within each phase of a system life-cycle. The impact of an undetected specification or design error within a project can appear to be of tremendous importance. The possibilities to perform verification or validation are either to make, when possible, a prototype of the real system and evaluate its appropriateness, or to build a model of the system and to analyze it. The most current approach is then to run some simulation; a complementary one is to handle formal reasoning without any execution. The paper develops a process for formal proof of properties of discrete state models. The main stress concerns the temporal evolution of the modeled systems. The method consists in representing the behavior of the system thanks to a state model, to translate it into a temporal logic framework and then to evaluate the existence of properties expressed as formulae. The first application is made on the classical FSM. It is then extended to the interpreted sequential machine, which is a state model able to consider any type of data. Moreover, using some graph theory results, it was possible to widen the proof of properties to the concept of "generic future".

http://ieeexplore.ieee.org/iel5/6569/17622/00814225.pdf

Janine Magnier

Papers

Formal Representation and Proof of the Interpreted Sequential Machine Model

Temporal and Functional Verification of a Symbolic Representation of Complex Systems

A process for improving multi-technology system high level design: modeling, verification and validation of complex optronic systems

Temporal logics for analyzing the behavior of systems

Extended temporal proof of properties using the interpreted sequential machine model