Post-doc position on Artificial Intelligence for flexible objects robotic handling and manipulation
Date de debut : 04-01-2021
Lieu : Palaiseau
Durée : 1 year (renewable once)
Salaire : depending on cursus
As part of CEA, CEA LIST institute is committed to technological innovation in digital systems. Its R&D activities - driven by major economic, societal and industrial challenges - encompass four main themes: factory of the future, cyber-physical systems, artificial intelligence and digital health. Within the CEA LIST, the Interactive Robotics Lab (SRI) investigate research on collaborative robots design and control, and apply it to different industrial and societal challenges.
MERGING (Manipulation Enhancement through Robotic Guidance and Intelligent Novel Grippers) is a collaborative research project funded by the European Commission under the Horizon 2020 Framework Programme (http://www.merging-project.eu/) coordinated by CEA LIST. It aims to provide manufacturers with a versatile, easy-to-use and low-cost robotized solution to handle flexible and fragile objects in industrial applications. By addressing challenges such as developing handling devices which are intelligent and universally dexterous, and making robots capable of manipulating soft products while controlling their movement and deformation, it will lead to disruptive innovations in many sectors.
Current robots are mostly used for industrial applications, where they handle rigid objects and interact with them in repetitive operations. But a big part of the European and global industry is comprised of sectors where soft and flexible items are manipulated. Such materials do not behave in an entirely predictable way, due to their high flexibility, and can be easily damaged. Grasping and manipulating them requires a gentle grasp, fine and adaptive control of movements that is beyond the current state of the art in robotics. The ambition of the MERGING project is to overcome these challenges and provide manufacturers with a turnkey robotic solution. Proof-of-concept studies will be carried out in three different sectors, in collaboration with the industrial partners of the project: fabric handling for lingerie manufacturing, technical fiber handling for composite manufacturing for the automotive industry, plastic pouches handling for the packaging in food industry.
CEA LIST (http://www-list.cea.fr/recherche-technologique/programmes-de-recherche/manufacturing-avance) offers a one-year (renewable) post-doc position on Artificial Intelligence applied to robot control. In the MERGING project context, it is intended the robot to learn and improve its behavior from its experience. Thus, deep learning techniques and other AI techniques will be investigated to improve the adaptability of the robot to the non-predictable object behavior and environment constraints in real time (i.e. during execution phases on the production line).
More precisely, the objectives of the post-doc are to develop, apply and integrate AI methods for robotics to:
- propose algorithms to estimate the operator intention in order to improve the control of the robot arm during flexible object co-manipulation operation. This control law adaptation has to consider human-robot as well as multi-robots system coordination;
- propose new algorithms to allow the robot improve its behavior (in terms of manipulation task efficiency) by its own experience, through proprio- and extero-ceptive sensor measurements;
- elaborate an algorithm able to propose automatically to a non-specialist operator, during the execution phase of a multiple-step industrial task, the best-fitted scheme intelligently chosen among a set of predefined assistance tools, with the aim to improve adaptability and flexibility to the task progress.
Applicants must have received a PhD in Artificial Intelligence ideally applied to robot control, and have experienced robot control implementation on real robots/test beds. They should have good programmation skills.
Aptitudes to English language are also necessary, in order to facilitate exchanges with the European project partners during project meetings, and also for scientific dissemination and publication.
Exceptionnal candidates with an engineer diploma in the domain and a first experience in a research environment will be also considered.
Institut CEA LIST / DIASI / SRI CEA Saclay - NanoInnov Bât. 863-Point Courrier n°178 F-91191 Gif-sur-Yvette Cedex - FRANCE
Post-doc at CEA requires candidates who should have received their PhD diploma at least before the beginning of the post-doctoral contract, or since two years at a maximum.
Post-doc salary at CEA depends on past diplomas and experience.
All application corresponding partly or totally to the position will be examined.
After a first selection on received documents, interviews will be organized. The chosen candidate will then have to fill in a detailled form in order to launch the internal recruitement procedure at CEA. The final acceptation of the chosen candidate will be validated only after receiving this last procedure conclusion from CEA recruitepment departement.