Teaching Autonomous Vehicles to Drive with Visible and Non-Visible Light Simulation | Unreal Fest 22

preview_player
Показать описание
Based on Unreal Engine, AURELION from dSPACE is a commercially available solution that enables car manufacturers to comprehensively test and validate their software and hardware for autonomous vehicles. In this talk, Caius Seiger, Product Manager for Sensor Simulation at dSPACE, explores the benefits of autonomous driving, reveals how autonomous cars work, and explains how AURELION integrates data from camera, radar, and LiDAR sensors for precise environment simulation.

Рекомендации по теме
Комментарии
Автор

Excellent presentation, great information and speaking. Thank you!

ABCD
Автор

2:47 - Fatality Rate due to Car Accidents in the US
5:13 - What is Autonomous Driving (SAE Level)
8:41 - Current State of Autonomous Driving
11:01 - How does an Autonomous Car work? Let's compare it to a Human Driver
14:46 - Video: Physics-based Sensor Simulation - AURELION
18:18 - From OpenSceneGraph to Unreal Engine
18:35 - A Cinema for Camera based Control Units
20:01 - Evolution from OpenSceneGraph to Unreal Engine
21:00 - Camera Simulation
28:55 - Radar Simulation
35:22 - Lidar Simulation
42:57 - Cloud Simulation - SIMPHERA
46:06 - Creating a Digital Twin
47:12 - Outlook: From UE4 to UE5

caiusseiger
Автор

Yeah, Pretty work. I am very interested in custom post-processing. It processes the intermediate results of rendering by calling CUDA's dynamic library in UE, which is simply wonderful. So I also want to try learning this. Do you have any information or demos explaining how to implement post-processing details by calling CUDA's dynamic library in UE?

JA-ocrb