I worked on this research activity between 2001 and 2010 in former laboratories ICTT and then LIESP. This research dealt with computer-aided hands-on laboratories (e-laboratories) in the more general context of E-Learning.



Remote Labs

With Hcene Benhmohamed, we have proposed a generic architecture for a remote-lab (RL) editing chain in order to offer tools common to all authors of RL training scenarios (real or virtual), usable whatever the discipline or the technological device, while taking advantage of existing tools (LMS, LCMS) for other more conceptual content. This work has gradually been the subject of several communications between 2002 and 2008 (see below). The common points (pedagogical objectives, pedagogical activities , communication between human actors, ...) between real and virtual RL pushed us to propose a chain indistinctly integrating these two forms of e-lab like the COLAB platform. The existing devices for online training are equipped with sufficient and interesting functionalities to be able to meet the needs specific to this type of training (learner monitoring, collaborative work, administration system, etc.). It therefore did not seem judicious to us to develop a new one, but on the contrary, to graft to these devices tools implementing additional functionalities, specific to RL. These tools must be sufficiently modular and generic to be able to be integrated and reused, we have had to study the standardization of these devices.

Scientific approach

We proposed the following approach:
  1. modeling of particular practical exercises, in four stages (observation of three situations of automatic practical work face-to-face, identification of invariable aspects from one lab to another, conceptualization then verification and validation by a fourth practical work)
  2. Construction of a RL model from the previous lab model by identifying invariable topics between face-to-face and remotely, taking into account the constraints (temporal, spatial, organizational, etc.) then extension (addition of additional functions) and ablation (removal of impossible functions) of the model as well as got
  3. validation of the teleTP model using 4 realizations (top-down and iterative approach): instantiation of the model in four particular RL, observation of the actual course of the training act, realization a return of uses in order to make possible corrections to the model then possible correction of this model depending on the results of the user feedback.

Generic model of automation laboratories

We have therefore developed a generic model around three key elements:
  • an IT environment providing:
    • the remote operation tools of a system in the broad sense of the term and which can typically be a model pedagogical, with a level of controllability and observability sufficient for the distance between the manipulated system and the learner hinder learning as little as possible,
    • an “engine” of training scenarios with online supervision and evaluation functionalities,
    • a means of communication (of the videoconference type) between learners and trainer,
  • a methodology helping trainers to distance their practical lab models via this device,
  • an authoring environment to edit scenarios without being a computer expert.
To gradually validate our proposals, we used and gradually improved a platform experimental teleTP which had been designed in 2000-2001, so just before my arrival at ICTT [LR] , in the framework of the collective project 1 Octopus. We were helped in 2001-2002 by other students as part of the project collective TIPY [P] . The students produced a first version of the work management and distribution environment remote practices and organized a user feedback campaign. The answers to the interviews were synthesized from the point of view of the educational objectives expected from the practical work, the role of the teachers in the session, the interventions of the teachers, the evaluation of the students, the resources teaching materials made available to students, the chronology of sessions, student activity, constraints

Generic E-Lab Edition Chain

This publishing chain is the first iteration of an ongoing project between 2001 and 2010. To be able to offer RL to trainees, it is necessary to offer tools covering the entire process of editing content from writing from the pedagogical scenario associated with a device to the rendering of the final report by the learners. This chain has an interest only if it offers possibilities of

  • reuse content in other scenarios;
  • reuse this content on other devices if it has been organized sufficiently independently of the pedagogical system from the start;
  • adapt these contents to similar devices (for example on a model having the same functionalities but a slightly different implementation: scopes, PLCs, software, ... from and/or brands and different models, etc.).

The life cycle that we have proposed includes three main stages :

  • the declaration of the various technological devices within a teleTP platform (called ELaMS);
  • editing and re-editing of “generic” scenarios (in the same family of devices with identical functions);
  • the use of these scenarios by tutors after adaptation to the target device.

Our contribution for this life cycle has been a platform integrating tools for E-Learning standards (LMS , LCMS) and a middleware offering them specific functionalities for RL This software is capable of exchanging data in the standard IMS-LD format. We used Coppercore (OUNL) as LMS. It was expected that version 2 of Moodle integrates this standard but this was ultimately not the case.
We have designed the ELaMS (Electronic Laboratory Management System) middleware, whose functionalities are to install and reference new training devices, open their access to trainees and tutors (depending on their own access rights, availability of devices and features required in the training scenario).
It was able to automatically direct a learner to a free device (once and for all or at each manipulation) using scheduling algorithms. An analysis of functional risks related to RL platforms was carried out using the FMEA method, to highlight the critical elements for which preventive and curative solutions must be put in place to ensure continuity of service. ELaMS was based on functional descriptions of the components and functions of each device, coded with ontologies. For this, we were inspired by the techniques of the Semantic Web by adopting a recent standard at the time: OWL, standardized by the W3C.
These ontologies were hosted on an ontology server (called OntoServ) and described the components and the functions they offer to RL actors. We created these ontologies via Protégé software (Stanford University), a royalty-free ontology editing software. They were public, posted on a web server.
For this ontology development work, I teamed up with Jacques Fayolle and Christophe Gravier LT2C, who were also working on RL with a more “low level” approach in the sense of “IT-Telecom”. We worked together on the structure of these ontologies so that they could work easily with the laboratory device remote tools they were developing.
ELaMS also integrated functionalities for managing and planning the use of shared resources (the training models). From a practical point of view, we had reused a free tool PhpScheduleIt. We had nevertheless looked into strategies for optimizing the planning of these resources by making the parallel with task scheduling in real-time systems. The design and implementation of the ELaMS tool are detailed in Hcene's Ph.D. report.

My publications on this topic


Here is the list of publications related to this research topic.