Background:

High-Definition (HD) maps provide centimeter-level geometric and semantic representations that enable precise localization and safe navigation for autonomous systems. However, most existing HDmapping pipelines rely on offline processing, requiring manual annotation and global optimization. Such approaches are unsuitable for dynamic or unstructured environments like sidewalks, where the scene continuously changes and map updates must occur in real time. Recent research on online HD-map construction has shown promising progress toward real-time, vectorized map generation from streaming LiDAR and camera data. Zhang et al. (2025) introduced MapExpert, a sparse map-element expert framework for efficient incremental HD-map generation, while StreamMapNet (2024) proposed a streaming architecture for continuous vectorized mapping. Lyu (2025) further provided a comprehensive survey summarizing the transition from offline, vehicle-centric mapping toward adaptive, online frameworks capable of continuous updates in openworld conditions. Building upon these advances, this thesis investigates online HD-mapping for humanoid sidewalk navigation, where narrow corridors, irregular curbs, and frequent dynamic obstacles demand real-time, semantically rich, and adaptive map representations beyond conventional driving scenarios.

Your Tasks:

• Conduct a comprehensive literature review on HD-map construction and online mapping frameworks, emphasizing methods for urban and pedestrian-scale environments.
• Investigate sidewalk-aware localization and mapping using LiDAR and camera data on humanoid robot platforms.
• Develop a real-time HD-mapping pipeline combining geometric mapping and semantic segmentation (e.g., curb, crosswalk, pedestrian zones).
• Evaluate the proposed pipeline using public datasets and real robot experiments.

Your Profile:

• Solid understanding of computer vision, robotics, and mapping.
• Hands-on experience with LiDAR and camera data processing (ROS 2 preferred).
• Familiarity with SLAM frameworks (e.g., LIO-SAM) and HD-map representations
• Independent, research-oriented mindset and the ability to explore open-ended problems.

Start date: Immediately

Duration: As per the applicable examination regulations.

If you are interested or have any questions regarding this thesis position, feel free to contact.