Background:

World models aim to predict how an environment evolves over time conditioned on an agent's observations and motion. While recent occupancy-based world models have shown  promising results in autonomous driving, they are typically designed for a single type of embodied agent. It remains unclear whether learned world dynamics can generalize across heterogeneous navigation embodiments operating in the same physical environment. This Master's thesis investigates Cross-Embodiment Occupancy World Models for Urban Navigation, focusing on three representative agents: road vehicles, UAVs, and sidewalk mobile robots. These embodiments operate in shared outdoor urban environments but differ substantially in viewpoint, motion space, speed, and interaction patterns. A simulation pipeline and synthetic dataset covering these three embodiments are already available. Building on this infrastructure, the thesis will focus on developing and evaluating occupancy-based world models that can transfer learned scene dynamics across embodiments. The primary goal is to study cross-embodiment future occupancy prediction, rather than end-to-end navigation or control.

Your Tasks:

• Conduct a literature review on occupancy world models, and cross-embodiment learning.
• Adapt and extend an existing occupancy world model [1] to support Car, UAV, and sidewalk robot data within a unified formulation.
• Evaluate single-embodiment, joint-training, and cross-embodiment generalization across the three agent types, and analyze which scene dynamics transfer across embodiments.
• Write your thesis; publication of outstanding results is highly encouraged.

Your Profile:

• Strong background in computer vision, deep learning, robotics, or related fields. 
• Motivation to work on open-ended research problems and conduct independent experiments. 
• Solid programming skills in Python and experience with PyTorch. 
• Good analytical skills and ability to communicate research ideas clearly in English.

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.