State of the Art analysis for Sim2Real approaches for Autonomous Driving

State of the art results in domain transfer have enhanced the use of game engines for testing autonomous driving functionalities. Recently published methods take sensor inputs (camera, LiDARs etc.) from simulators and transfer them to real world domain thereby evaluating their performance.

Within the research project BEINTELLI we develop a scalable software solution for our vehicle, edge and cloud infrastructure. Domain transfer plays a significant role when transferring various results to a real test vehicle. This master thesis is focused on implementation of state-of-the-art methods for domain transfer and evaluate their performance on our research test vehicle.

Prerequisites

  • Good Python programming skills, optional experiences with ROS
  • Knowledge about deep learning & generative models
  • Experience with deep learning frameworks (Pytorch, Tensorflow)
  • PS: Concrete tasks will be formulated based on interview & your interests

Literature

  • A Sim2real method based on DDQN for training a self-driving scale car,Mathematical Foundations of Computing,2019-12-20,Qi Zhang,Tao Du,Changzheng Tian.
  • Chisari, E., 2021. Learning from Simulation, Racing in Reality: Sim2Real Methods for Autonomous Racing. [online] Research-collection.ethz.ch.

Application

We are looking forward to receiving your PDF application. Please send your application with the following documents:

  • Current transcript of records
  • Curriculum vitae in tabular form