Research Field: Computer science Engineering Mathematics
Researcher Profile: First Stage Researcher (R1)
Application Deadline: 26/09/2022 23:59 - Europe/Brussels
Location: France › COMPIEGNE
Type Of Contract: Temporary
Job Status: Full-time
Hours Per Week: 35
Offer Starting Date: 01/10/2022
This thesis subject will be carried out in close collaboration between the
SyRI team of Heudiasyc and the ACENTAURI project team of INRIA Sophia
Antipolis and this within the framework of the ANR ANNAPOLIS project.
The proposed thesis subject aims to make the decision-making of intelligent
vehicles (IV) even more robust and safe during the complete delegation of
driving phases, and this in highly dynamic and constrained urban environments.
In this context, the VI can face multiple visual occlusions and unpredictable
behaviors of the moving entities around.
The main objective of this thesis is to propose a global probabilistic multi-
controller architecture (P-MCA) with a reliable risk assessment and
management system. The main challenges are to have the safest movement of the
VIs, even in complex environments/situations, but also to ensure the fluidity
of the trajectories of these vehicles (thereby guaranteeing the comfort of
the passengers). To achieve these objectives, mainly inspired by the work
given in Adouane 17, Iberraken 18 a new global Probabilistic Multi-
Controller Architecture, adapted to navigation in urban areas, must be
developed and embedded in each VI.
This architecture also aims to jointly use adaptive and model-predictive
controls (based on both steering/braking capabilities and the desired
response of the VI in the short to medium term) to generate safe trajectories
even in large diversity of driving conditions, uncertain and unexpected
events/situations. These modifications will inevitably lead to proposing an
appropriate stochastic control law with robust properties Dahmane 18, such
as the one based on the stochastic MPC (Model Predictive Control)
Mesbab016. MPC uses models to predict future developments within a
particular time horizon. A promising approach is also to explore the
potentialities of the Model Predictive Path Integral (MPPI) Williams 16,
which is a sample-based MPC that shows good results in autonomous navigation
under difficult conditions Philippe 19.
Assessing and managing vehicle risk is an important part of the intended
overall control architecture. In the literature, many methods have been used
as decision methods Schubert 12. Probabilistic decision making aims to
make the best continuous decisions in constrained and uncertain environments.
To do this, a robust and operational Markov decision-making process based on
the Multi-level Bayesian Decision Making Network (MB-DMN), as illustrated in
Iberraken 19b, will be developed. This type of process makes it possible
to have reliable retrospections on the consequence of the actions taken by the
VI according to the expected trajectories of the surrounding entities. Its
purpose is to minimize the risk of collisions of the VI when it is confronted
with dangerous and/or unexpected situations. The extension of the MB-DMN
together with appropriate augmented perception and better metrics to
characterize the probabilistic behaviors and trajectories of surrounding
entities will enhance the reliability of the targeted decision process.
Adouane 17, L. Adouane, Reactive versus cognitive vehicle navigation based
on optimal local and global PELC. Robotics and Autonomous Systems (RAS), ,
volume 88, pp. 51–70, February 2017, DOI 10.1016/j.robot.2016.11.006
Iberraken 19a, D. Iberraken, L. Adouane and D.Dieumet, "Multi-Controller
Architecture for Reliable Autonomous Vehicle Navigation: Combination of Model-
Driven and Data-Driven Formalization", Workshop FRCA-IAV, IEEE 2019 IEEE
Intelligent Vehicles Symposium.
Iberraken 19b, D. Iberraken, L. Adouane, and D. Dieumet, "Reliable Risk
Management for Autonomous Vehicles based on Sequential Bayesian Decision
Networks and Dynamic Inter-Vehicular Assessment", IEEE 2019 IEEE Intelligent
Ben-Lakhal 19, N.M Ben-Lakhal, L. Adouane, O. Nasri, and J. Ben Hadj
Slama, Risk Management for Intelligent Vehicles based on interval analysis of
TTC, 10th IFAC Symposium on Intelligent Autonomous Vehicles (IAV'19), 3-5
July 2019, Gdansk-Poland.
Maurer 16, Maurer, M., Gerdes, J.C., Lenz, B., Winner, H, .Autonomous
Driving: Technical, Legal and Social Aspects, Springer, 978-3-662-48845-4,
Philippe 19, C. Philippe, L. Adouane, B. Thuilot, A. Tsourdos and H-S.
Shin, Risk and Comfort Management for Multi-Vehicle Navigation using a
Flexible and Robust Cascade Control Architecture, European Conf. on Mobile
Robotics, Paris-France, 2017.
Schubert 12, Schubert, R. Evaluating the utility of driving: Toward
automated decision making under uncertainty. IEEE Transactions on Intelligent
Transportation Systems, 13(1), 354-364, 2012.
Mesbab016, A. Mesbah, "Stochastic Model Predictive Control: An Overview
and Perspectives for Future Research," in IEEE Control Systems Magazine, vol.
36, no. 6, pp. 30-44, Dec. 2016, doi: 10.1109/MCS.2016.2602087.
Williams2016, G. Williams, P. Drews, B. Goldfain, J. M. Rehg and E. A.
Theodorou, "Aggressive driving with model predictive path integral control,"
2016 IEEE International Conference on Robotics and Automation (ICRA), 2016,
pp. 1433-1440, doi: 10.1109/ICRA.2016.7487277.
Web site for additional job details
https: // emploi.cnrs.fr/Offres/Doctorant/UMR7253-LOUADO-001/Default.aspx
Required Research Experiences
YEARS OF RESEARCH EXPERIENCE
YEARS OF RESEARCH EXPERIENCE
YEARS OF RESEARCH EXPERIENCE
REQUIRED EDUCATION LEVEL
Engineering: Master Degree or equivalent
Computer science: Master Degree or equivalent
Mathematics: Master Degree or equivalent
Department: Heuristique et Diagnostic des Systèmes Complexes
Organisation Type: Public Research Institution
Website: https:// www. hds.utc.fr