Deep Learning: Robustifying HD Map Perception against Localization and Map Label Noise

Forschungsthema/Bereich

Autonomous Driving, High Definition Maps, Deep Learning

Typ der Abschlussarbeit

Bachelor / Master

Startzeitpunkt

05.01.2026

Bewerbungsschluss

01.05.2026

Dauer der Arbeit

4-6 months

Beschreibung

Current state-of-the-art map construction methods such as MapTRv2 use sen-
sor data (360° surround view camera setup and LiDAR) to construct high defini-
tion (HD) maps. These methods consists of two components the map encoder
and the map decoder. The map encoder extracts features from the sensor data
and transform them into a Bird’s Eye View (BEV) representation, whereas the
map decoder derives a map in polyline representation using transformer-based
architectures from those BEV features. Prior work has demonstrated that the
performance of these models is significantly degraded by localization noise (aris-
ing from inaccurate ego-poses) and noisy HD map labels (caused by temporal
degradation or inaccurate annotations). While numerous methods for label noise
robustness have been developed in the field of Computer Vision (e.g., MentorNet,
Co-Teaching, or robust loss functions), their application and rigorous evaluation
within the geometrically sensitive domain of HD Map Perception remain largely
unexplored.

The goal of this thesis is to adapt and implement various robustification methods
to an HD map perception framework (e.g., MapTRv2) to successfully overcome
performance degradation caused by localization noise and noisy labels. The
work will be conducted using the Argoverse 2 dataset.

The proposed thesis consists of the following parts:

+ Literature research about Robustification against noise in Deep Learning
+ Familiarization with the MapTRv2 framework and the Argoverse 2 dataset
+ Implementation of various robustification methods against localization noise
and noisy labels

I am happy to answer any questions you might have. Feel free to ask for an
appointment or directly ask at my office!

Link to the announcement:

• https://www.mrt.kit.edu/z/download/studip/Deep_Learning_Robustifying_HD_Map_Perception_against_Localization_and_Map_Label_Noise.pdf

Voraussetzung

Voraussetzungen an Studierende

• Expericence with Neural Networks in Deep Learning context
• Experience with PyTorch, NumPy and Matplotlib
• Motivation and independent work style with the interest learning new things

Studiengangsbereiche

• Ingenieurwissenschaften
  Elektrotechnik & Informationstechnik
  Informatik
  Maschinenbau
  Mechatronik & Informationstechnik
  Mechanical Engineering
  Computer Science
  Electrical Engineering and Information Technology
  Mechatronics and Information Technology
• Naturwissenschaften und Technik
  Computational and Data Sience

Betreuung

Titel, Vorname, Name

Jonas, Merkert

Organisationseinheit

Institut für Mess- und Regelungstechnik (MRT)

E-Mail Adresse

jonas.merkert@kit.edu

Link zur eigenen Homepage/Personenseite

Website

Bewerbung per E-Mail

Bewerbungsunterlagen

• Lebenslauf
• Notenauszug
• Immatrikulationsbescheinigung

E-Mail Adresse für die Bewerbung
Senden Sie die oben genannten Bewerbungsunterlagen bitte per Mail an jonas.merkert@kit.edu


Zurück