Student Projects
Please contact team members working in your area of interest and apply with your CV and transcripts. In case you have project ideas related to any of these projects, take the opportunity and propose your own project!
We can also support Master Theses and or exchange semesters at renowned universities around the world. Please contact us in case of interest!
Studies on Mechatronics: We offer students to conduct their Studies on Mechatronics at our lab. In general, we recommend to do the Studies on Mechatronics in combination with the Bachelor Thesis, either as prepartory work the semester before or as extended study in parallel.
Contact
Mobile Robotics Lab
Leonhardstrasse 21
LEE J 206
8092
Zurich
Switzerland
Estimating Human-Object Interaction from Egocentric Videos for Humanoid Whole-Body Control
This project focuses on reconstructing 3D human-object interactions directly from uncurated egocentric video by estimating object trajectories, collision meshes, and contact points to lay the foundation for downstream humanoid control.
Keywords
Human-Object Interaction, 4D Reconstruction, Learning from Demonstrations
Labels
Semester Project , Master Thesis
Description
Contact Details
More information
Published since: 2026-08-03 , Earliest start: 2026-08-03 , Latest end: 2027-04-30
Organization Mobile Robotics Lab
Hosts Brandes Alex
Topics Information, Computing and Communication Sciences , Engineering and Technology
Humanoid Retargeting for Whole-Body Loco-Manipulation
This project aims to bridge the human-robot embodiment gap in humanoid control by unifying motion retargeting and RL tracking into an adaptive framework that replaces rigid joint-position matching with flexible, rich motion references.
Keywords
Human2Robot Retargeting, RL, Motion Tracking, Humanoids
Labels
Master Thesis
Description
Contact Details
More information
Published since: 2026-08-03 , Earliest start: 2026-09-01 , Latest end: 2027-04-30
Organization Mobile Robotics Lab
Hosts Brandes Alex , Hansson Alexander
Topics Information, Computing and Communication Sciences , Engineering and Technology
LiDAR-Visual-Inertial Odometry with a Unified Representation
LiDAR-Visual-Inertial odometry approaches [1-3] aim to overcome the limitations of the individual sensing modalities by estimating a pose from heterogenous measurements. Lidar-inertial odometry often diverges in environments with degenerate geometric structures and visual-inertial odometry can diverge in environments with uniform texture. Many existing lidar-visual-inertial odometry approaches use independent lidar-inertial and visual-inertial pipelines [2-3] to compute odometry estimates that are combined in a joint optimisation to obtain a single pose estimate. These approaches are able to obtain a robust pose estimate in degenerate environments but often underperform lidar-inertial or visual-inertial methods in non-degenerate scenarios due to the complexity of maintaining and combining odometry estimates from multiple representations.
Labels
Semester Project , Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2026-07-02 , Earliest start: 2026-07-01
Organization Mobile Robotics Lab
Hosts Le Gentil Cédric , Barbas Sebastian
Topics Information, Computing and Communication Sciences , Engineering and Technology
Scalable Humanoid Control through Quantized Latent Action Representations
Traditional approaches to humanoid robotics often require training and maintaining N separate policies for N distinct skills (e.g., one policy for walking, one for sitting, one for standing up). This becomes highly unscalable as the robot's repertoire grows. Whole-body motion tracking policies alleviate this problem by training a motion tracking policy on a large repertoire of robot motions. However, simple per-frame tracking policies require a high-level policy which generates reference motions. This project aims to replace isolated skills with a single, unified controller. By learning a Latent Action Model (LAM) that compresses multi-frame human motion into low-dimensional, quantized macro-commands, we can train a single low-level policy on the Unitree G1 to track a wide variety of behaviors dictated entirely by the latent space.
Keywords
Humanoid, Control, Robotics, AI, Reinforcement Learning, Motion Tracking
Labels
Master Thesis , ETH Zurich (ETHZ)
Goal
Contact Details
More information
Published since: 2026-06-21 , Earliest start: 2026-09-06
Organization Mobile Robotics Lab
Hosts Näf Joshua , Hansson Alexander
Topics Information, Computing and Communication Sciences , Engineering and Technology