Efficient transmission of 3D PointCloud data to safeguard autonomous vehicles

Institut

Lehrstuhl für Fahrzeugtechnik

Typ

Semesterarbeit /

Inhalt

 

Beschreibung

Automated vehicles are not yet fully operational in road traffic. However, teleoperation makes it possible to support the automated driving function in complex situations such as roadworks. Currently, teleoperation mainly uses video streams to transmit information about the surrounding situation. However, the camera-driven perception of the environment leads to a reduced "telepresence."

Leading companies in the field of autonomous driving are increasingly focussing on the display of 3D information using HD maps. However, the perception of the situation is not trivial despite the 3D visualization. One hypothesis is that these 3D visualizations can be improved with the help of PointClouds data. However, due to the amount of data, sending PointCloud data over the network is not trivial. This term paper aims to select and implement such a transmission method.

Ideally, the results of the research will be integrated into the chair's existing teleoperation software

The selection of the method can be chosen according to interest, e.g., projection methods in combination with classic video transmission, machine learning-based approaches, as well as the transmission of representations rendered on the vehicle side.

The following skills are required to complete the course:

• Literature research on the selected technique
• Basic knowledge of algorithms and data formats
• Solid knowledge of C++, testing can also be carried out in Python by arrangement
• Ideally, experience with 3D data formats from PointClouds and LiDAR sensors and experience with video transmission

Voraussetzungen

The following literature can serve as a first point of contact:

• RealTime Streaming Point Cloud Compression for 3D LiDAR Sensor Using U-Net
• Point Cloud Compression for Efficient Data Broadcasting: A Performance Comparison
• Graphics Library Transmission Format 
• Draco 

The paper should document the individual work steps in a clear form.

If you are interested or have any questions, please send an e-mail at niklas.krauss@tum.de with your CV and current transcript of records, thank you very much! 

Verwendete Technologien

Transmission, ROS2, PointCloud Compression, C++,

Tags

FTM Studienarbeit, FTM Krauss, FTM AV, FTM AV Safe Operation, FTM Informatik, FTM Teleoperation

Möglicher Beginn

Kontakt

Niklas Krauß
Raum: 3507
Tel.: +49172 1736882
niklas.krausstum.de

Ausschreibung