07-11, 15:00–15:10 (Europe/Amsterdam), While Loop (4.2)
The talk focuses on integrating Julia in a demonstrator car (Ford mondeo) to perform perception tasks (lidar pointcloud segmentation) in real-time.
A special focus is set on the Phyton - Julia interface and how both languages can work well together.
Future autonomous vehicles need to process a million LIDAR points in real time. Julia is a perfect fit for these perception tasks including ground segmentation, creation of a semantic occupancy grid or object detection.
The talk shows how to use Julia in the development of ML algorithms and integration process in a real car. Special focus will be at the algorithmic development process (compared to PyTorch), experience with the training phase, integration process in C++ / Python environments, parallelism and runtime issues.
The results presented in this talk have been created within the EU project AI4CSM (ai4csm.eu) (ECSEL Joint Undertaking program).
Gerhard Dorn is a senior scientist at Virtual Vehicle Research GmbH in Austria with a background in mathematics and computational physics.
Within the project AI4CSM Gerhard works on LIDAR perception and the integration into the test vehicle Ford Mondeo.
Gerhard is an active member of the 🎈 Pluto.jl and ⚡ PlutoUI.jl development community and likes to organize workshops as community manager.