A Lightweight Machine Learning Pipeline for LiDAR-simulation
This work addresses the domain gap in LiDAR simulation for autonomous driving testing, offering a more efficient alternative to expensive physics-based methods, though it is incremental as it builds on existing image-to-image translation techniques.
The paper tackles the problem of unrealistic LiDAR simulation in autonomous driving by proposing a lightweight approach that learns real sensor behavior from test drive data and applies it to virtual domains, achieving sufficient generalization from real to synthetic images without complex physics simulation.
Virtual testing is a crucial task to ensure safety in autonomous driving, and sensor simulation is an important task in this domain. Most current LiDAR simulations are very simplistic and are mainly used to perform initial tests, while the majority of insights are gathered on the road. In this paper, we propose a lightweight approach for more realistic LiDAR simulation that learns a real sensor's behavior from test drive data and transforms this to the virtual domain. The central idea is to cast the simulation into an image-to-image translation problem. We train our pix2pix based architecture on two real world data sets, namely the popular KITTI data set and the Audi Autonomous Driving Dataset which provide both, RGB and LiDAR images. We apply this network on synthetic renderings and show that it generalizes sufficiently from real images to simulated images. This strategy enables to skip the sensor-specific, expensive and complex LiDAR physics simulation in our synthetic world and avoids oversimplification and a large domain-gap through the clean synthetic environment.