Section: New Results

Smart*: design, multi-modeling and co-simulation and supervision of mobile CPS/IoT

Participants : Laurent Ciarletta [contact] , Ye-Qiong Song, Yannick Presse, Julien Vaubourg, Emmanuel Nataf, Petro Aksonenko, Virgile Dauge, Louis Viard, Florian Greff, Virginie Galtier, Thomas Paris.

Vincent Chevrier (former Maia team, Dep 5, LORIA) is a collaborator and the correspondant for the MS4SG/MECSYCO project, as well as Christine Bourjot (former MAIA team, Dep 5, LORIA).

Sylvain Contassot-Vivier (Dep 3, Loria) is a collaborator on the Grone project and is directing Virgile Daugé with Laurent Ciarletta.

Pierre-Etienne Moreau is a collaborator on the CEOS project and is directing Louis Viard with Laurent Ciarletta.

Virginie Galtier from CentraleSupélec is now a member of the Loria laboratory and will integrate the future Simbiot team (Systems of Interactive aMBient Intelligent ObjecTs).

In Pervasive or Ubiquitous Computing, a growing number of communicating/computing devices are collaborating to provide users with enhanced and ubiquitous services in a seamless way.

These systems, embedded in the fabric of our daily lives, are complex: numerous interconnected and heterogeneous entities are exhibiting a global behavior impossible to forecast by merely observing individual properties. Firstly, users physical interactions and behaviors have to be considered. They are influenced and influence the environment. Secondly, the potential multiplicity and heterogeneity of devices, services, communication protocols, and the constant mobility and reorganization also need to be addressed. Our research in this field is going towards both closing the loop between humans and systems, physical and computing systems, and taming the complexity, using multi-modeling (to combine the best of each domain specific model) and co-simulation (to design, develop and evaluate) as part of a global conceptual and practical toolbox.

We proposed the AA4MM meta-model [45] that solves the core challenges of multimodeling and simulation coupling in an homogeneous perspective. In AA4MM, we chose a multi-agent point of view: a multi-model is a society of models; each model corresponds to an agent and coupling relationships correspond to interaction between agents. In the MECSYCO-NG (formerly MS4SG, Multi Simulation for Smart Grids) project which involves some members of the former MAIA team, Madynes and EDF R&D on smart-grid simulation, we developed a proof of concepts for a smart-appartment case that serves as a basis for building up use cases, and we have worked on some specific cases provided by our industrial partner. We also collaborated with researchers from the Green UL laboratory.

In 2017 we worked on the following research topics:

• Overall assessment and evaluation of complex systems.

• Cyber Physical Systems and Smart *.

  We have continued the design and implementation of the Aetournos platform at Loria which will be part of the Creativ'Lab. The collective movements of a flock of flying communicating robots / UAVs, evolving in potentially perturbed environment constitutes a good example of a Cyber Physical System. Several projects have started during the last part of 2017. One of the emerging topic in this area is the safety of Mobile IoT/CPS with regards to their environment and users.

  • The Grone (Interreg) project involves partners from the 4 countries of the Grande Région (Centrale Supélec, LIST, Univ Luxembourg, Univ Liège, Fraunhofer IZFP to name a few). The main goal is to develop UAV based solution for the surveillance of industrial and agricultural sites and the exploration of GPS denied and underground environments. A PhD has been started in March 2017 (Systèmes cyber-physiques autonomes et communicants en milieux hostiles. Application à l'exploration par robots mobiles. Virgile Daugé).

  • and the CEOS (FUI22) project involving high profile companies (Thales C&S, EDF R&D, ENEDIS and Aéroport de Lyon) as well as academic partners (AOSTE2 Inria, ESIEE) and collaborating SMEs (RT@W, ADCIS, Alerion). This project focuses on the safety of UAV based monitoring solutions for OIV (Opérateurs d'Intéret Vital) infrastructures. A PhD has been started in November 2017 (Environnement de développement et d'analyse de propriétés pour des systèmes cyber-physiques mobiles. Louis Viard).

  The work on Software Defined Real-Time Mesh networks (Florian Greff's PhD CIFRE with Thales R&T) has given many results as he plans to defend his work in march 2018 [15], [39], [14].

  On more specific subject of innovative sensors for mobile and interactive IoT, a collaborative project with the KPI (Ukraine) university has been started with a projected PhD (Méthodes optimisées de calibration, d'alignement et algorithmes d'attitude avancés pour les systèmes de navigation inertiels fixés, Petro Aksonenko). Several papers have been published [9] [44] in 2017.

• (Very Serious) Gaming: Starburst Gaming. During some exploratory work, we have seen the potential of these Pervasive Computing ressources in the (Very Serious) Gaming area which led us to the Starburst Computing SATT projects in 2016 and 2017. A spin off has been founded in 2017 that is getting the licences for the resulting IP (the software is under the APP process at the time of this writing). Starburst is already involved in a AMI project with the Globlinz game studio and the lab and has officially been accepted in novembre 2017 and will be operational in 2018.

• Smart *: MS4SG / MECSYCO-NG has given us the opportunity to link simulations tools with a strong focus on FMI (Functional Mockup Interface) and network simulators (NS3/Omnet++). We have so far successfully applied our solution to the simulation of smart apartment complex and to combine the electrical and networking part of a Smart Grid. The AA4MM software is now MECSYCO and has seen constant improvements in 2017 thanks to the ressources provided by the MECSYCO-NG project in collaboration with EDF R&D (http://www.mecsyco.com), and the work of Thomas Paris and Julien Vaubourg.

  Starting from domain specific and heterogenous models and simulators, the MECSYCO suite allows for multi systems integration at several levels: conceptual, formal and software. A couple of visualization tools have been developed as proof of concepts both at run-time and post-mortem.

The technical report [43] has been extended into a journal paper under revision for a publication in 2018.