Hands-On Deep Learning: Optimization of Multi-Class Motion Prediction for Autonomous Driving

Institute

Lehrstuhl für Fahrzeugtechnik

Type

Semester Thesis / Master's Thesis /

Content

 

Description

Motion prediction of surrounding objects is a key requirement for autonomous driving. However, state-of-the-art approaches face a fundamental challenge: the motion patterns of different classes diverge strongly. Pedestrians may stop abruptly, cyclists combine pedestrian-like agility with vehicle-like dynamics, and vehicles typically follow traffic rules but can still vary widely in speed and maneuvers. Capturing all these behaviors within a single model is highly demanding, as class-specific nuances are often averaged out.

In this thesis, an existing joint tracking and prediction approach is to be further optimized to enable robust and accurate multi-class motion prediction. To achieve this, literature research will be conducted, and the existing model will be refined, with a particular focus on exploring class-specific prediction heads as a promising strategy for improving prediction performance.

Possibility of a publication in case of excellent work.

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