Automatic Trucking Darmstadt

Daimler Truck & TU Darmstadt conduct research for road safety

Together with Daimler Truck we are working on assisted and automated driving functions for trucks. One part of this collaboration focuses on Machine Learning based Free Space Detection and another one emphasizes on driver assistance for challenging truck-specific maneuvers such as roundabouts.

Research Topic: Supported Driving through Roundabouts

The construction of roundabouts at a later stage, particularly in urban areas, often results in undersized roundabouts due to the limited available space. Furthermore, the time pressure and the occasional lack of experience or unawareness of drivers contribute to the occurrence of minor accidents. Nevertheless, these incidents frequently result in significant consequences and substantial costs.

To address this issue, this project is analyzing map data and developing a trajectory planner. In the future this should assist the driver in avoiding collisions with the infrastructure when driving through roundabouts. During the course of this project, the entire pipeline for the development of the roundabout assistant, as well as the system architecture of such a system, will be examined.

The focus is on planning a collision-free trajectory through a roundabout. To this end, a simulation pipeline will be set up.

Bezier-modified and therefore collision-free trajectory on a recognized trafficked area, for a simulative scenario.
Bezier-modified and therefore collision-free trajectory on a recognized trafficked area, for a simulative scenario.

This research topic is being investigated by Jan Staab.

Research Topic: Sensor-based Free Space Detection

Truck accidents often occur because drivers misjudge the drivable space around the vehicle, resulting in collisions with road infrastructure. These incidents can result in expensive repairs or even the need to scrap vehicles, driving up costs for freight transportation companies.

The project aims to detect and calculate the available space around the vehicle to improve safety for truck drivers and other road users. This research focuses on developing a free space detection function using various automotive sensors, including cameras, lidars, and radars.

For that, different relevant use cases will be considered and the function's requirements will be derived. The goal is to help drivers identify potential obstacles on the road, such as lost cargo, bridges, and other road users, thereby improving overall road safety.

How a possible Free Space Representation could look like.
How a possible Free Space Representation could look like.

This research topic is being investigated by Cristian Martinez.