ORGANISATION/COMPANYUniversité Clermont Auvergne
RESEARCH FIELDComputer science
RESEARCHER PROFILEFirst Stage Researcher (R1)
APPLICATION DEADLINE28/06/2020 00:00 - Europe/Brussels
LOCATIONFrance › AUBIERE
TYPE OF CONTRACTTemporary
HOURS PER WEEK35 H
OFFER STARTING DATE01/10/2020
IS THE JOB RELATED TO STAFF POSITION WITHIN A RESEARCH INFRASTRUCTURE?Yes
Design by AI of innovative multi-physical systems integrating smart materials.
Yuri Lapusta, Professor.
Institut Pascal (UMR 6602 - UCA/CNRS/SIGMA), M3G axis (Mécanique, Génie
mécanique, Génie civil, Génie industriel).
Email and phone:
+33 4 73 28 80 49
Frédéric Chapelle, Associate professor.
Abstract (up to 10 lines):
The objective of this thesis is to develop a deep learning and multi-criteria
optimisation method for multi-physical systems integrating smart active materials.
The optimisation algorithm is bioinspired and should be used for the
synthesis of mechanical architectures, while deep learning should enable to
learn by reinforcement the optimal trajectories of a series of representative
tasks. The hybridisation of these two algorithms can be achieved by a sequential
and/or synchronous approach, to be studied. The mechanical architecture
is of a modular nature, each module reflecting a type of smart material
and its properties. The systems thus designed can be physically developed
We are looking for a master's degree student motivated by Artificial Intelligence,
innovative materials, active systems. The ideal profile includes a first
experience in AI programming.
Deep learning, Bio-inspired multi-criteria optimisation, Smart materials,
Multiphysics active systems.
Description (up to 1 page):
AI has been used to perform material modelling , structure/mechanism
optimisation , , image analysis , automatic programming . The recent
evolutions in deep learning reinforce the interaction capabilities of the
developed systems with their environment . At the same time, the development
of smart active materials reacting to multi-physical stimuli (electricity,
magnetic field, pressure, heat, etc.) ,  has led to the design of innovative
architectures for these systems .
These evolutions make it necessary to develop AI algorithms specifically
adapted to the design of multi-physical systems and to the definition of the
tasks they enable to perform. The interest is, for example, to obtain a global
design approach taking into account, during the same optimisation cycle, the
possibilities of physical design and trajectory shape generation for the
achievement of the desired tasks. Such an approach would make it possible
to preserve the interactions between architectural synthesis and task implementation
synthesis during optimisation. The targeted tasks will include follow-
the-leader trajectories, where the system must progress along a path
without deviating from it, as well as scanning a volume, useful for domestic
and medical applications.
The solution will be based on the construction of a hybrid method integrating
two algorithms known for their efficiency in multi-criteria optimisation (bioinspired
such as evolutionary algorithm or particle swarm) and learning by
reinforcement (neural networks). The nature of hybridisation will be studied
during the thesis with determination of the degree of sequential/parallel
synchronisation between the two algorithms. The system architectures will
be modular, each module considering a type of smart material, its shape, and
its physical properties.
Innovative design: multiphysics system + task
aggregating candidate solutions
Artificial neural networks
References (up to ½ page):
 A. Mendizabal, P. Márquez-Neila, and S. Cotin, ‘Simulation of hyperelastic materials
in real-time using deep learning’, Medical image analysis, vol. 59, p.
 F. Chapelle and P. Bidaud, ‘Evaluation functions synthesis for optimal design of
hyper-redundant robotic systems’, IFToMM Mechanism and Machine Theory,
vol. 41, no. 10, pp. 1196–1212, 2006.
 R. J. Alattas, S. Patel, and T. M. Sobh, ‘Evolutionary modular robotics: Survey and
analysis’, Journal of Intelligent & Robotic Systems, vol. 95, no. 3–4, pp. 815–828,
 N. Wang and D.-Y. Yeung, ‘Learning a deep compact image representation for
visual tracking’, in Advances in neural information processing systems, 2013, pp.
 F. Chapelle and P. Bidaud, ‘Closed form solutions for inverse kinematics approximation
of general 6R manipulators’, IFToMM Mechanism and Machine Theory,
vol. 39, no. 3, pp. 323–338, 2004.
 N. Sünderhauf et al., ‘The limits and potentials of deep learning for robotics’,
The International Journal of Robotics Research, vol. 37, no. 4–5, pp. 405–420,
 H. Yuan, F. Chapelle, J.-C. Fauroux, and X. Balandraud, ‘Concept for a 3D-printed
soft rotary actuator driven by a shape-memory alloy’, Smart Materials and
Structures, vol. 27, no. 5, p. 055005, 2018.
 V. Loboda, A. Sheveleva, F. Chapelle, and Y. Lapusta, ‘A dielectric breakdown
model for an electrode along an interface between two piezoelectric materials’,
Engineering Fracture Mechanics, vol. 224, p. 106809, 2020.
 D. Trivedi, C. D. Rahn, W. M. Kier, and I. D. Walker, ‘Soft robotics: Biological inspiration,
state of the art, and future research’, Applied Bionics and Biomechanics,
vol. 5, no. 3, pp. 99–117, Dec. 2008.
How to candidate?
Send an email with CV, cover letter, and grades for the available years of
study (including the first semester of the current year) to the following addresses:
REQUIRED EDUCATION LEVELOther: Master Degree or equivalent
EURAXESS offer ID: 528421
The responsibility for the jobs published on this website, including the job description, lies entirely with the publishing institutions. The application is handled uniquely by the employer, who is also fully responsible for the recruitment and selection processes.
Please contact firstname.lastname@example.org if you wish to download all jobs in XML.