TED – 3D Object Detection with Transformers for Autonomous Driving

Institute

Lehrstuhl für Fahrzeugtechnik

Type

Master's Thesis /

Content

experimental / theoretical /  

Description

Autonomous vehicles rely on an accurate and complete understanding of their surroundings. This means knowing where agents are (object detection) and how they will move in the future (motion prediction). Object detection contributes to this goal by detecting agents of different semantic classes around the ego vehicle. It highly influences downstream motion prediction, as upstream errors are propagated through the software stack, which leads to imperfect prediction inputs. Prediction is thus directly coupled to the detection performance.

The goal of this thesis is to develop a novel 3D object detection algorithm, which uses LiDAR point cloud and additional contextual information as inputs. Specifically, in a previous work, trajectories of agents (taken from a motion prediction module) were used as additional information for the detector. This approach should be further developed as it led to promising results. Specifically, the approach should be extended to transformer-based object detector architectures and evaluated in a representative way

---

Work packages:

• Literature review on 3D object detectors and trajectory feedback.

• Development of a robust evaluation pipeline

• Development and optimization of the trajectory enhanced object detector

• In-depth evaluation.

 

If you are interested, please apply with your CV and transcript of records. 

Requirements

Requirements:

• Very good programming skills in Python.

• High personal motivation and independent working style.

• Very good language proficiency in German, English or French.

Software

Python

Tags

FTM AV, FTM AV Perception, FTM Stratil, FTM Informatik

Possible start

sofort

Contact

Loïc Stratil, M.Sc.
Room: MW 3508
Phone: +49.89.289.15898
loic.stratiltum.de

Announcement