Scientific direction Development of key enabling technologies
Transfer of knowledge to industry

PostDocs : selection by topics

See all positions

Formalization of the area of responsibility of the actors of the electricity market

DPACA (CTReg)

Autre DPACA

01-06-2020

PsD-DRT-20-0074

bruno.robisson@cea.fr

The CEA is currently developing a simulation tool which models the energy exchanges between the actors of the electricity market but which models, in addition, the exchanges of information between those actors. The first results of this work show that, for some new energy exchange schemes, 'indirect' interactions between actors may appear and may cause financial damage (for example, the failure of a source of production of one actor may impact the income of another). Thus, the borders which clearly delimited until now the areas of responsibility of each actors could be brought to blur and their areas of responsibility could "overla". The candidate will be responsible for: - Formally define the area of responsibility of an actor in the electricity market, - Model the interactions, including 'indirect' ones, that may appear between these actors, - Apply formal proof techniques (such as 'model-checking') to detect overlaps in areas of responsibility, - Define the conditions of exchange between the actors which would guarantee the non-recovery of the areas of responsibility.

Download the offer (.zip)

High precision robotic manipulation with reinforcement learning and Sim2Real

Département Intelligence Ambiante et Systèmes Interactifs (LIST)

Laboratoire Vision et Apprentissage pour l'analyse de scènes

01-09-2020

PsD-DRT-20-0082

jaonary.rabarisoa@cea.fr

High precision robotic assembly that handles high product variability is a key part of an agile and a flexible manufacturing automation system. To date however, most of the existing systems are difficult to scale with product variability since they need precise models of the environment dynamics in order to be efficient. This information is not always easy to get. Reinforcement learning based methods can be of interest in this situation. They do not rely on the environment dynamics and only need sample data from the system to learn a new manipulation skill. The main caveat is the efficiency of the data generation process. In this post-doc, we propose to investigate the use of reinforcement learning based algorithms to solve high precision robotic assembly tasks. To handle the problem of sample generation we leverage the use of simulators and adopt a sim2real approach. The goal is to build a system than can solve tasks such as those proposed in the World Robot Challenge and tasks that the CEA's industrial partners will provide.

Download the offer (.zip)

Application of ontology and knowledge engineering to complex system engineering

Département Ingénierie Logiciels et Systèmes (LIST)

Labo. ingénierie des langages exécutables et optimisation

01-06-2019

PsD-DRT-19-0088

florian.noyrit@cea.fr

Model-Based System Engineering relies on using various formal descriptions of the system to make prediction, analysis, automation, simulation... However, these descriptions are mostly distributed across heterogeneous silos. The analysis and exploitation of the information are confined to their silos and thereby miss the big picture. The crosscutting insights remain hidden. To overcome this problem, ontologies and knowledge engineering techniques provide desirable solutions that have been acknowledged by academic works. These techniques and paradigm notably help in giving access to a complete digital twin of the system thanks to their federation capabilities, in making sense to the information by embedding it with existing formal knowledge and in exploring and uncovering inconsistencies thanks to reasoning capabilities. The objective of this work will be to propose an approach that gives access to a complete digital twin federated with knowledge engineering technologies. The opportunities and limits of the approach will be evaluated on industrial use cases.

Download the offer (.zip)

Characterization of X-ray emitting radionuclides - Application to reactor dosimetry

Département Métrologie Instrumentation et Information (LIST)

Laboratoire de Métrologie de l'Activité

01-09-2019

PsD-DRT-19-0090

marie-christine.lepy@cea.fr

The activity measurement of X-ray emitting radionuclides in the energy range below 100 keV encounters several difficulties that limit the accuracy of the result. These include the difficulty of calibrating detector performance and, in general, the significant uncertainties associated with emission intensities X. In addition, the self-absorption effects of X-rays in standard sources or samples lead to important corrections that must be controlled. Among the important applications of X-emitter measurement, reactor dosimetry, which makes it possible to determine the neutron fluence received during irradiation and to characterize its spectrum, is based on the analysis of the activity of irradiated dosimeters. These are made of pure metals or alloys of perfectly known compositions, some of which are activated or fissioned by neutrons. For example, reactions 93Nb(n,n')93Nbm and 103Rh(n,n')103Rhm are of prime importance for reactor dosimetry and are particularly interesting for characterizing neutron fluxes around 1 MeV. The proposed work follows a thesis that identified several areas for improvement in dosimeter measurement that will need to be implemented, including : - improvement of radionuclide X-ray emission data used as standard for calibration (133Ba, 152Eu, etc.) to establish a consistent set of data; - validation of corrective coefficients due to the presence of impurities during dosimeter irradiation; - evaluation and publication of the decay scheme of 103Pd and 103mRh; - implementation of a new method of performance calibration using monochromatic radiation.

Download the offer (.zip)

Advanced Runtime Assertion Checking of C Programs

Département Ingénierie Logiciels et Systèmes (LIST)

Laboratoire pour la Sûreté du Logiciel

01-10-2020

PsD-DRT-20-0094

Julien.Signoles@cea.fr

Frama-C is a framework for code analysis of C code. E-ACSL is the Frama-C plug-in dedicated to runtime assertion checking. It converts a C program extended with formal annotations into a new C program which checks the validity of annotations at runtime: it behaves in the same way than the input program if all its annotations are valid, or (by default) the program execution stops whenever one annotation is violated. One key feature of E-ACSL is the expressiveness of its specification language which allows the user to describe powerful safety and security properties. Another key feature is the efficiency of the generated code which relies on a custom memory library and dedicated static analyses. Still, many challenging research questions are opened to go beyond the state of the art of runtime assertion checkers and improve E-ACSL significantly. They include (but are not limited to): - runtime assertion checking of axiomatic definitions by relying on synthesis techniques - runtime assertion checking of localized properties that refer to several program points - runtime assertion checking of frame conditions and data dependency properties - runtime assertion checking of properties over real numbers - static analysis for monitor optimisation In the context of the H2020 European project ENSURESEC (2020-2022) that aims at protecting e-commerce by monitoring, the hired postdoc researcher will collaborate with other engineers and researchers at LSL and possibly in other labs, in order to address several of these challenges. She or he will design, formalize and implement innovative solutions, and prove their soundness.

Download the offer (.zip)

Design and control of a multimodal multi-fingered gripper for force and vision-guided robotic manipulation

Département Intelligence Ambiante et Systèmes Interactifs (LIST)

Laboratoire d'Architecture des Systèmes Robotiques

01-01-2021

PsD-DRT-20-0096

mathieu.grossard@cea.fr

The current offer focuses on the design and control approaches of functionally integrated multi-fingered grippers to be used for force- and vision-guided robotic manipulation tasks. The to-be-designed grasping tool will take advantage of multimodal sensing capabilities (embedding tactile sensors, mounted in a meaningful way within the mechatronic device, together with a self-sensing actuation apparatus) combined with a vision system to achieve autonomously fine force-controlled manipulation tasks (such as dexterous in-hand manipulation, pick-and place of rigid and flexible objects, component assembly, etc.). To do so, the previously designed multimodal system will be exploited for learning-by-demonstration tasks (relying on Human/Robot co-manipulation phase) to build a learnt database combining tactile, force and vision cues for a series of objects with various characteristics (in terms of geometry, shape, texture, mass, etc.). This database would first cover known objects, before being extended to objects in various poses and then to unknown objects using learning-based strategies.

Download the offer (.zip)

32 Results found (Page 4 of 6)
first   previous  2 - 3 - 4 - 5 - 6  next   last

See all positions