PhD position in dependable robots, LAAS-CNRS, Toulouse
PhD Position in Robotics & Dependability at LAAS-CNRS, Toulouse, France
Robot Navigation Testing in Virtual Worlds
LAAS-CNRS seeks to hire an outstanding doctoral researcher in robotics and dependability in Toulouse, France. The successful candidate will collaborate with the RIS (Robotics) and TSF (Dependability) teams to develop a systematic method for simulation-based testing of autonomous robot software.
An autonomous robot should perform a variety of tasks without any human supervision, in various, complex and uncertain environments. In order to deploy such systems in human or critical environments, rigorous validation is of paramount importance. The tests, for safety and cost constraints, must first be realized using simulation in virtual worlds. MORSE (Modular OpenRobots Simulation Engine) is a simulator developed at LAAS-CNRS to serve that purpose. It provides the right level of simulation (perception/action) to validate the execution of complex plans. The objective of the PhD work is to develop a test method that exploits such simulation facilities to place the robotic system in a wide variety of operational situations and observe how it faces them.
The method will be based on the automated generation of virtual worlds, like worlds generated in video games. The main contribution will be on
how to synthesize the virtual worlds in such a way that they exhibit the safety-r! elevant corner cases, and ensure adequate coverage of the functionalities deployed on the robot. The method will be applied to navigation functionalities developed in GenoM, which are deployed on real mobile robots at LAAS. A demonstrator will be implemented using the MORSE simulator.
Desired profile and role
-----------------------
The candidate is expected to:
- Have a Master in computer science,
- Have some background and interest in robotics and/or software validation,
- Be fluent in English (written and spoken).
The main tasks of the PhD student will be to:
- Get acquainted with the MORSE robots simulator and the GenoM3 module generator tool
- Propose a framework for MORSE world generation and selection according to test criteria
- Define plans and scripts which will be tested in simulation
- Conduct a case study and evaluate the proposed solution
How to apply
Send a resume that provides detailed information about your education
(including grades of the last three university years), motivations,
scientific background and interests, and 2 references to:
helene.waeselynck@laas.fr, jeremie.guiochet@laas.fr,
felix.ingrand@laas.fr
A more detailed description of the PhD subject (in French) can be found
here: http://homepages.laas.fr/guiochet/telecharge/testnav.pdf
Lab, and research teams:
www.laas.fr
www.laas.fr/TSF/
www.laas.fr/RIS