Détail de l'offre

[Master thesis/internship] Comparison methodology of dynamic occupancy grids for autonomous driving

Offre du 15/11/2022

Comparison methodology of dynamic occupancy grids for autonomous driving
Supervisors: P. Koch, J-B. Horel, C. Laugier

1 Context
Obstacle detection is a key component of autonomous vehicles. The ability to detect dynamic obstacles, along with estimating their velocity, allows to track them in order to predict future states, hence generate safe commands. A common solution to this problem is to generate dynamic occupancy grids. However, quantitatively evaluating their performance and validating their correct functioning is still a challenge to be solved. This internship will be focused on finding new solutions for the dynamic grids performance evaluation and to use them to compare two existing methods: an in-house algorithm [1] and a state of the art solution [2]. This work will be done in the Chroma team (1) at Inria Grenoble Rhone-Alpes research center, within the "Plateforme de Recherche et d'Investissement pour la Sûreté et la Sécurité de la Mobilité Autonome" (2) project.

(1) https://team.inria.fr/chroma
(2) https://prissma.fr

2 Objectives
The candidate will be tasked to study and evaluate dynamic occupancy grid based methods, with a particular focus on two existing approaches. To do so, the starting point will be to design suitable metrics, in order to evaluate the perception algorithm at a ”fine” level. Those metrics should be developed with validation [3] in mind, to be re-used in automated testing. The internship is composed of three main tasks: first, a state of the art of robotic perception metrics, with a focus on occupancy grids, then implementation of metrics toolbox, and finally a thorough analysis of the obtained results.

3 Requirements
• Master 2 in robotics, computer science, image processing
• Python, C++, LaTeX, Linux/Ubuntu
• Bonus: OpenCV, ROS1, numpy

4 Additional Information
• The intern will have the opportunity to perform tests on Chroma self-driving vehicle [4].
• The internship duration is 5 or 6 months, ideally starting from February/March 2023.
• To apply, please send your resume and cover letter to: pierrick.koch@inria.fr


38330 Montbonnot-Saint-Martin


Type de Contrat

  • Stage
  • Fonction

    • Ingénieur.e
    • Diplômes

      • Bac +5 (M2, Ingénieur, etc.)
      • Expérience

        • Débutant
        • Compétences Techniques

          • Vision
          • Langages Informatiques

            • C, C++
            • Python
            • LaTeX
            • Outils et Environnements

              • ROS
              • OpenCV
              • Linux