Projects and Theses
Utilizing human body for ambient electromagnetic energy harvesting
The goal of the project is to develop wearable devices, for use in environmental electromagnetic energy recovery based on human body application.
Keywords
Flexible electronics, electromagnetic energy harvesting
Labels
Semester Project , Bachelor Thesis , Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-04-23 , Earliest start: 2025-05-31
Organization Biomedical and Mobile Health Technology Lab
Hosts Li Yuanlong
Topics Engineering and Technology
Gaussian Avatar Reconstruction from Single Image
In this project, you are going to work with a state-of-the-art deep learning approach and generative models for building an efficient system to directly reconstruct a 3D animatable avatar from a single image. Feel free to contact me for more details.
Keywords
3D Gaussian Avatar, Diffusion Model, 3D from single image
Labels
IDEA League Student Grant (IDL) , Semester Project , Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2025-04-23 , Earliest start: 2025-05-01 , Latest end: 2025-12-15
Organization Advanced Interactive Technologies
Hosts Dong Zijian
Topics Information, Computing and Communication Sciences
Mechanophores for advanced wearable strain and pressure sensors
The goal of the project is to synthesize and characterize a number of small molecules capable of acting as mechanophore addition to various polymers. These polymers would then be used as wearable strain or pressure sensors.
Keywords
mechanophore, polymer, wearable, sensor, color, strain, pressure
Labels
Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-04-22 , Earliest start: 2024-09-01 , Latest end: 2025-09-01
Organization Biomedical and Mobile Health Technology Lab
Hosts Shokurov Aleksandr
Topics Engineering and Technology , Chemistry
Point-of-Care Sensor for Urinary Iodine
The goal of the project is to develop a cheap and disposable sensor capable of determination of iodine levels in human urine for early diagnostic purposes.
Keywords
electrochemistry, iodine, nutrition, health, point of care
Labels
Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-04-22 , Earliest start: 2025-01-01 , Latest end: 2025-10-01
Organization Biomedical and Mobile Health Technology Lab
Hosts Shokurov Aleksandr
Topics Medical and Health Sciences , Engineering and Technology , Chemistry
Metal oxide-based textile logic electronics
The goal of the project is to develop metal oxide-based logic electronics, for use in rectification and/or memristor application.
Keywords
Metal oxide, smart textile, wearable electronics, logic electronics, electrochemistry
Labels
Semester Project , Bachelor Thesis , Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-04-21 , Earliest start: 2025-05-31 , Latest end: 2025-11-30
Organization Biomedical and Mobile Health Technology Lab
Hosts Li Yuanlong
Topics Engineering and Technology
Ideal fixation technique for unstable fragility fractures of the sacrum
Recently, the management of fragility fractures of the pelvis has gained increasing attention. Rommens and Hofmann have introduced a comprehensive classification of these fracture patterns with increasing instability from grade I to IV[1]. U- and H-shaped fractures of the sacrum (FFP IVb) are the fractures with the highest instability due to a complete spinopelvic dissociation[2]. Non-operative treatment may be associated with impaired walking abilities, chronic pain, and the potential loss of independence. However, different treatment options are still controversially debated. The aim of surgical treatment includes sufficient fracture stability for immediate full weight bearing and good pain control postoperatively. A new surgical treatment algorithm was developed[3], suggesting the surgical fixation of FFP IVb fractures with two transiliac-transsacral screws (TI-TSS) in the first intact corridor cranial to the horizontal fracture pattern or lumbopelvic fracture fixation combined with one TI-TSS. However, the biomechanical stability of the different fixation techniques has not been evaluated.
Labels
Semester Project , Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-04-16 , Earliest start: 2025-05-01 , Latest end: 2026-06-30
Organization Snedeker Group / Laboratory for Orthopaedic Biomechanics
Hosts Fasser Marie-Rosa
Topics Medical and Health Sciences , Engineering and Technology
Tissue Engineering Approaches to Study Tendon Injury, Disease, and Therapy
Join a dynamic research team at the intersection of biomechanics, tissue engineering, and cell biology. This project offers hands-on training in state-of-the-art methods to investigate how tendon tissue responds to injury, disease processes, and mechanical stimulation during exercise-based therapy.
Keywords
Tendon biology, tissue engineering, mechanobiology, cell culture, microscopy, regenerative medicine, exercise therapy, inflammation, ECM remodeling
Labels
Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-04-15 , Earliest start: 2025-05-01 , Latest end: 2026-12-31
Organization Snedeker Group / Laboratory for Orthopaedic Biomechanics
Hosts Snedeker Jess, Prof.
Topics Engineering and Technology
Experimental and Numerical Investigation of Direction-Dependent Flow Resistance in Engineered Geometries
Controlling fluid flow is essential in various natural and engineering systems, with geometry playing a fundamental role in shaping fluid behavior. However, the interaction between geometry and flow behavior remains a complex phenomenon, primarily governed by the flow regime and fluid material properties. Certain geometries, whether naturally occurring or engineered, induce direction-dependent flow resistance, causing variations in velocity and flow rate in opposite directions. One well-known example of such engineered geometries is the Tesla valve—a passive device without moving parts, designed to create asymmetric flow resistance, particularly at high Reynolds numbers. This structure acts like a fluidic diode, offering greater resistance to flow in one direction by generating turbulent vortices and flow separations while allowing smoother movement in the opposite direction. This effect is quantified by diodicity, which represents the ratio of pressure drop in the reverse direction to that in the forward direction, providing a measure of the valve's asymmetric resistance. However, this direction dependence is limited at lower velocities. We have designed two sets of geometries that effectively induce directional flow resistance within high and low fluid flow velocities. This Master’s thesis project aims to experimentally investigate the impact of different flow obstruction designs on direction-dependent resistance in rectangular channels and semicircular arc segments. The student will, together with their direct supervisor, design and construct an experimental setup for the reliable measurement of flow and diodicity. This project offers an excellent opportunity to gain expertise in fluid dynamics, experimental testing, numerical modeling, and additive manufacturing, with applications in biomedical systems. Students with a background in mechanical engineering, fluid dynamics, or related fields are encouraged to apply. Prior experience with COMSOL Multiphysics is beneficial but not mandatory.
Labels
Master Thesis
Contact Details
More information
Published since: 2025-04-15 , Earliest start: 2025-06-01 , Latest end: 2025-12-01
Organization Musculoskeletal Biomechanics
Hosts Mosayebi Mahdieh
Topics Engineering and Technology
Research Assistant with data collection,cleaning,processing and programming skills
We are looking for a research assistant who is skilled at data collection, cleaning, matching and programming. Please see details in the attachment.
Labels
Student Assistant / HiWi
More information
Published since: 2025-04-13 , Earliest start: 2025-04-13 , Latest end: 2025-05-15
Organization Chair of Strategic Management and Innovation
Hosts Liu Chang
Topics Information, Computing and Communication Sciences
Hardware Design Internship in Brain Imaging
Join us in revolutionizing brain imaging technologies and make it accessible for everyday use. Functional near-infrared spectroscopy (fNIRS) is an emerging technology that enables cost-effective and precise brain measurements, helping to improve neurotherapies and brain health.
Keywords
3D-printing, injection molding, design, brain imaging, neuro, wearables, health, startup
Labels
Internship
Description
Goal
Tasks
Your Profile
Contact Details
More information
Published since: 2025-04-09 , Earliest start: 2025-04-10 , Latest end: 2025-06-26
Organization Rehabilitation Engineering Lab
Hosts Wyser Dominik
Topics Medical and Health Sciences , Information, Computing and Communication Sciences , Engineering and Technology
PCB design for neuromorphic vision on nano-drones
In this project, we aim to develop a novel PCB integrating a powerful PULP chip, i.e., the GAP9, and event-based sensor, the Prophesee Genx320, and a RGB camera, the Himax HB0360 to enable multi modal AI-driven perception aboard nano-drones
Keywords
PCB design, nano-drones, robotics, event-cameras, neuromorphic computing, embedded devices, ultra-low-power
Labels
Semester Project , Bachelor Thesis , Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-04-07 , Earliest start: 2025-04-07 , Latest end: 2025-12-01
Organization Digital Circuits and Systems (Benini)
Hosts Lamberti Lorenzo
Topics Engineering and Technology
Continual Learning and Domain Adaptation Techniques for a Waste Monitoring System on an Ocean Cleanup Vessel
This thesis develops an automated onboard waste quantification system for a maritime waste collection vessel, leveraging computer vision with continual learning and domain adaptation to replace manual counting of floating waste. Evaluated under real-world maritime conditions, the system aims to improve waste management in the South East Asian Sea.
Keywords
Computer Vision, Continual Learning, Field Testing
Labels
Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-03-26 , Earliest start: 2025-05-01 , Latest end: 2025-12-31
Organization Robotic Systems Lab
Hosts Stolle Jonas , Elbir Emre
Topics Engineering and Technology
Master Thesis / Project - SENSEI: Sensor Teaching in Multi-Activity classification from Video and Wearables for Wheelchair Users
In this project, we focus on continuous and quantitative monitoring of activities of daily living (ADL) in SCI individuals with the goal of identifying cardiovascular events and PI-related risk behaviors. ADLs specific to SCI patients and their lifestyles shall be discussed and narrowed down in the scope of this work, therefore an autonomous camera-based system is proposed to classify ADLs. The Current work builds on a previous project where a SlowFast network [1] was trained to identify SCI-specific classes and we aim to further improve the classification and temporal resolution for transferring to wearables' time-series data.
Keywords
Computer vision, activity classification, video processing, Deep Learning, ADL, soft-labelling, probabilistic networks
Labels
Semester Project , Course Project , Internship , Bachelor Thesis , Master Thesis , ETH for Development (ETH4D) (ETHZ) , ETH Zurich (ETHZ)
Project Background
Your Task
Your Benefits
Your Profile
Contact Details
More information
Published since: 2025-03-25 , Earliest start: 2025-05-01 , Latest end: 2026-02-28
Applications limited to EPFL - Ecole Polytechnique Fédérale de Lausanne , ETH Zurich , Zurich University of the Arts , Wyss Translational Center Zurich , University of Zurich , Zurich University of Applied Sciences , CERN , CSEM - Centre Suisse d'Electronique et Microtechnique , Department of Quantitative Biomedicine , Lucerne University of Applied Sciences and Arts , Institute for Research in Biomedicine , IBM Research Zurich Lab , University of St. Gallen , University of Lucerne , University of Lausanne , University of Geneva , University of Fribourg , University of Berne , University of Basel , Université de Neuchâtel , Swiss Institute of Bioinformatics , Swiss National Science Foundation , Swiss Federal Institute for Forest, Snow and Landscape Research , Institute of Robotics and Intelligent Systems D-MAVT , TU Berlin , TU Darmstadt , TU Dresden , RWTH Aachen University , Technische Universität München , Technische Universität Hamburg , Max Planck Society , University of Oxford , University of Leeds , University of Cambridge , UCL - University College London , National Institute for Medical Research , Imperial College London , Royal College of Art , Empa , Università della Svizzera italiana , Hochschulmedizin Zürich , Hong Kong University of Science and Technology , University of Washington , Tokyo Institute of Technology , The University of Tokyo
Organization Sensory-Motor Systems Lab
Hosts Paez Diego, Dr. , Paez Diego, Dr. , Paez Diego, Dr.
Topics Medical and Health Sciences , Information, Computing and Communication Sciences , Behavioural and Cognitive Sciences
Extending Functional Scene Graphs to Include Articulated Object States
While traditional [1] and functional [2] scene graphs are capable of capturing the spatial relationships and functional interactions between objects and spaces, they encode each object as static, with fixed geometry. In this project, we aim to enable the estimation of the state of articulated objects and include it in the functional scene graph.
Keywords
scene understanding, scene graph, exploration
Labels
Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-03-25 , Earliest start: 2025-03-25
Applications limited to ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne
Organization Computer Vision and Geometry Group
Hosts Bauer Zuria, Dr. , Trisovic Jelena , Zurbrügg René
Topics Information, Computing and Communication Sciences , Engineering and Technology
Master Thesis: Development of a Customized Knee Orthosis for Osteoarthritis
Osteoarthritis (OA) presents a significant challenge in healthcare, necessitating innovative solutions to alleviate pain, enhance mobility. This thesis documents the research and development journey of an OA knee orthosis within the Spinal Cord and Artificial Intelligence Lab (SCAI-Lab) at ETH Zurich. This thesis is a close collaboration between the ORTHO-TEAM Group and the SCAI-Lab at ETH Zurich. The collaboration offers a unique exchange of expertise and resources between industry and academia. Together, we aim to make meaningful progress in the field of and empower students to make valuable contributions to their academic pursuits.
Keywords
Osteo Arthritis, Orthosis, Biomechanics, AI, Medical Data, Healthcare
Labels
Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2025-03-25 , Earliest start: 2025-04-15 , Latest end: 2026-01-31
Applications limited to ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne , Empa , University of Basel , University of Berne , Zurich University of Applied Sciences , Università della Svizzera italiana , Hochschulmedizin Zürich , Lucerne University of Applied Sciences and Arts , Institute for Research in Biomedicine , CSEM - Centre Suisse d'Electronique et Microtechnique
Organization Spinal Cord Injury & Artificial Intelligence Lab
Hosts Paez Diego, Dr. , Paez Diego, Dr.
Topics Medical and Health Sciences , Engineering and Technology
Development of a Heterocellular Human Bone Organoid for Precision Medicine and Treatment
Our goal is to establish a heterocellular 3D printed bone organoid model comprising all major bone cell types (osteoblasts, osteocytes, osteoclasts) to recapitulate bone remodeling units in an in vitro system. The organoids will be produced with the human cells, as they could represent human pathophysiology better than animal models, and eventually could replace them. These in vitro models could be used in the advancement of next-generation personalised treatment strategies. Our tools are different kinds of 3D bioprinting platforms, bio-ink formulations, hydrogels, mol-bioassays, and time-lapsed image processing of micro-CT scans.
Keywords
3D printing, bone organoids, co-culture, bioreactor, hydrogels, drug testing
Labels
Semester Project , Internship , Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2025-03-24 , Earliest start: 2022-08-01 , Latest end: 2025-11-30
Organization Müller Group / Laboratory for Bone Biomechanics
Hosts Steffi Chris
Topics Engineering and Technology , Biology
Exploring the Mechanoregulation of Bone Regeneration
In over 100 years, the remarkable ability of bone to adapt to its mechanical environment has been a source of scientific fascination. Bone regeneration has been shown to be highly dependent on the mechanical environment at the fracture site. It has been demonstrated that mechanical stimuli can either accelerate or impede regeneration. Despite the fundamental importance of the mechanical environment in influencing bone regeneration, the molecular mechanisms underlying this phenomenon are complex and poorly understood.
Keywords
Bone, Mechanobiology, Spatial transcriptomics, Gene expression, Finite element modelling, Image processing
Labels
Semester Project , Internship , Bachelor Thesis , Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-03-23 , Earliest start: 2024-11-01 , Latest end: 2025-08-31
Organization Müller Group / Laboratory for Bone Biomechanics
Hosts Mathavan Neashan
Topics Medical and Health Sciences , Engineering and Technology
Brain plasticity after congenital limb loss
Previous studies have demonstrated that following the loss of an upper limb, the deprived hand territory of the somatosensory cortex becomes responsive to afferent input of intact body parts (e.g., the face). It is hypothesised that this remapping of body parts is partially driven by adaptive behaviours, whereby the body part most often used to compensate for the missing limb is remapped into the cortical hand area. We are seeking motivated and independent students to assist with participant recruitment, MRI scanning, behavioural testing and fMRI data analysis.
Keywords
reorganisation, biomechanics, congenital limb loss, neuroscience, movement analysis, MRI, recruitment
Labels
Semester Project , Internship , Lab Practice , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2025-03-19 , Earliest start: 2025-03-20 , Latest end: 2025-08-01
Organization Neural Control of Movement Lab
Hosts Howell Paige
Topics Medical and Health Sciences , Information, Computing and Communication Sciences , Engineering and Technology , Behavioural and Cognitive Sciences
Event-based feature detection for highly dynamic tracking
Event cameras are an exciting new technology enabling sensing of highly dynamic content over a broad range of illumination conditions. The present thesis explores novel, sparse, event-driven paradigms for detecting structure and motion patterns in raw event streams.
Keywords
Event camera, neuromorphic sensing, feature detection, computer vision
Labels
Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-03-13 , Earliest start: 2025-03-17
Organization Robotic Systems Lab
Hosts Kneip Laurent
Topics Engineering and Technology
Fast, change-aware map-based camera tracking
Experiment with Gaussian Splatting based map representations for highly efficient camera tracking and simultaneous change detection and map updating. Apply to different exteroceptive sensing modalities.
Keywords
Localization, Camera Tracking, Gaussian Splatting, Change detection
Labels
Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-03-13 , Earliest start: 2025-03-17
Organization Robotic Systems Lab
Hosts Kneip Laurent
Topics Engineering and Technology
Soft object reconstruction
This project consists of reconstructing soft object along with their appearance, geometry, and physical properties from image data for inclusion in reinforcement learning frameworks for manipulation tasks.
Keywords
Computer Vision, Structure from Motion, Image-based Reconstruction, Physics-based Reconstruction
Labels
Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-03-13 , Earliest start: 2025-03-17
Organization Robotic Systems Lab
Hosts Kneip Laurent
Topics Engineering and Technology
Reconstruction from online videos taken in the wild
Push the limits of arbitrary online video reconstruction by combining the most recent, prior-supported real-time Simultaneous Localization And Mapping (SLAM) methods with automatic supervision techniques.
Keywords
Computer Vision, 3D Reconstruction, SLAM
Labels
Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-03-13 , Earliest start: 2025-03-17
Organization Robotic Systems Lab
Hosts Kneip Laurent
Topics Engineering and Technology
Computationally Efficient Neural Networks
Computing, time, and energy requirements of recent neural networks have demonstrated dramatic increase over time, impacting on their applicability in real-world contexts. The present thesis explores novel ways of implementing neural network implementations that will substantially reduce their computational complexity and thus energy footprint.
Keywords
AI, CNNs, transformers, network implementation
Labels
Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-03-12 , Earliest start: 2025-03-17
Organization Robotic Systems Lab
Hosts Kneip Laurent
Topics Engineering and Technology
Generalist Excavator Transformer
We want to develop a generalist digging agent that is able to do multiple tasks, such as digging and moving loose soil, and/or control multiple excavators. We plan to use decision transformers, trained on offline data, to accomplish these tasks.
Keywords
Offline reinforcement learning, transformers, autonomous excavation
Labels
Semester Project , Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-03-11 , Earliest start: 2025-03-01 , Latest end: 2025-08-31
Organization Robotic Systems Lab
Hosts Werner Lennart , Egli Pascal Arturo , Terenzi Lorenzo , Nan Fang , Zhang Weixuan
Topics Information, Computing and Communication Sciences
Exploring the 3D Mineralization Behavior in Material-Induced Osteoinduction Through a Multiscale Micro-CT Imaging Approach
The project aims at investigating material-induced osteoinduction using the available mouse model of orthotopic or ectopic bone graft substitute (BGS) application. Through the 3D-3D registration of ex vivo and in vivo multiscale micro-CT images, crucial 3D mineralization behavior of the BGS can be investigated.
Keywords
Femur, Bone Graft Substitute, Critical Size Defect, Osteoinduction, in vivo, micro-CT, 3D-3D Image Registration, Image Analysis, Image Processing, Python, Computational
Labels
Semester Project , Bachelor Thesis , Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-03-11 , Earliest start: 2025-04-01 , Latest end: 2026-01-31
Organization Müller Group / Laboratory for Bone Biomechanics
Hosts Lindenmann Sara
Topics Medical and Health Sciences , Engineering and Technology
Differential Particle Simulation for Robotics
This project focuses on applying differential particle-based simulation to address challenges in simulating real-world robotic tasks involving interactions with fluids, granular materials, and soft objects. Leveraging the differentiability of simulations, the project aims to enhance simulation accuracy with limited real-world data and explore learning robotic control using first-order gradient information.
Labels
Semester Project , Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-03-10 , Earliest start: 2025-01-01 , Latest end: 2025-12-31
Applications limited to ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne
Organization Robotic Systems Lab
Hosts Nan Fang , Ma Hao
Topics Engineering and Technology
Wearable kirigami antenna for motion monitoring
The aim of the project is to develop a simple method for fabrication of kirigami-inspired laser-cut or molded antennas on flexible substrates. This technology will enable advancements in wearable electronics for wireless communication and sensing applications.
Keywords
wearable, flexible electronics, kirigami, laser cutting, 3D printing, antenna design, conductivity, wireless communication
Labels
Semester Project , Bachelor Thesis
Description
Goal
Contact Details
More information
Published since: 2025-03-09 , Earliest start: 2025-03-24 , Latest end: 2026-08-31
Organization Biomedical and Mobile Health Technology Lab
Hosts Kateb Pierre
Topics Engineering and Technology
Master Thesis: Contact force evaluation of robotic endoscopic system based on Series Elastic Actuation
In the BIROMED-Lab we have been developing an endoscopic system for safer neurosurgeries with inspiration from human finger anatomy. Its two degrees of freedom allow the endoscope to investigate areas of the brain that would be inaccessible with standard rigid endoscopes. Thanks to springs in the transmission between the motors and the movable endoscope tip, the interaction forces between the instrument and the brain tissue can be reduced. Furthermore the interaction forces can be estimated by measuring the deflection of the spring. To make the telemanipulation of the endoscope safer and more intuitive for the surgeon, force feedback was also implemented.
Keywords
Robotic surgery, Neurosurgery, Telemanipulation, Haptic feedback, Robotic endoscope
Labels
Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-03-06 , Earliest start: 2025-03-01
Organization Bio-Inspired RObots for MEDicine-Laboratory (BIROMED-Lab)
Hosts Ettori Sara Lisa Margherita , Gerig Nicolas, Dr. , Sommerhalder Michael
Topics Engineering and Technology
Novel Winch Control for Robotic Climbing
While legged robots have demonstrated impressive locomotion performance in structured environments, challenges persist in navigating steep natural terrain and loose, granular soil. These challenges extend to extraterrestrial environments and are relevant to future lunar, martian, and asteroidal missions. In order to explore the most extreme terrains, a novel winch system has been developed for the ANYmal robot platform. The winch could potentially be used as a fail-safe device to prevent falls during unassisted traverses of steep terrain, as well as an added driven degree of freedom for assisted ascending and descending of terrain too steep for unassisted traversal. The goal of this project is to develop control policies that utilize this new hardware and enable further climbing robot research.
Keywords
Robot, Space, Climbing, Winch, Control
Labels
Semester Project , Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-03-05 , Earliest start: 2024-10-07
Organization Robotic Systems Lab
Hosts Vogel Dylan
Topics Information, Computing and Communication Sciences , Engineering and Technology
Beyond Value Functions: Stable Robot Learning with Monte-Carlo GRPO
Robotics is dominated by on-policy reinforcement learning: the paradigm of training a robot controller by iteratively interacting with the environment and maximizing some objective. A crucial idea to make this work is the Advantage Function. On each policy update, algorithms typically sum up the gradient log probabilities of all actions taken in the robot simulation. The advantage function increases or decreases the probabilities of these taken actions by comparing their “goodness” versus a baseline. Current advantage estimation methods use a value function to aggregate robot experience and hence decrease variance. This improves sample efficiency at the cost of introducing some bias. Stably training large language models via reinforcement learning is well-known to be a challenging task. A line of recent work [1, 2] has used Group-Relative Policy Optimization (GRPO) to achieve this feat. In GRPO, a series of answers are generated for each query-answer pair. The advantage is calculated based on a given answer being better than the average answer to the query. In this formulation, no value function is required. Can we adapt GRPO towards robot learning? Value Functions are known to cause issues in training stability [3] and a result in biased advantage estimates [4]. We are in the age of GPU-accelerated RL [5], training policies by simulating thousands of robot instances simultaneously. This makes a new monte-carlo (MC) approach towards RL timely, feasible and appealing. In this project, the student will be tasked to investigate the limitations of value-function based advantage estimation. Using GRPO as a starting point, the student will then develop MC-based algorithms that use the GPU’s parallel simulation capabilities for stable RL training for unbiased variance reduction while maintaining a competitive wall-clock time.
Keywords
Robot Learning, Reinforcement Learning, Monte Carlo RL, GRPO, Advantage Estimation
Labels
Semester Project , Bachelor Thesis , Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-03-05
Organization Robotic Systems Lab
Hosts Klemm Victor
Topics Information, Computing and Communication Sciences , Engineering and Technology , Behavioural and Cognitive Sciences
Volumetric Bucket-Fill Estimation
Gravis Robotics is an ETH spinoff from the Robotic Systems Lab (RSL) working on the automation of heavy machinery (https://gravisrobotics.com/). In this project, you will be working with the Gravis team to develop a perceptive bucket-fill estimation system. You will conduct your project at Gravis under joint supervision from RSL.
Keywords
Autonomous Excavation
Labels
Semester Project , Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-02-28 , Earliest start: 2025-01-01 , Latest end: 2026-01-01
Organization Robotic Systems Lab
Hosts Egli Pascal Arturo
Topics Engineering and Technology
Master Thesis: Vibro-tactile feedback in ventricle puncturing during External Ventricular Drain (EVD) procedure
EVD is a common procedure in Neurosurgery, nevertheless its placement is non-ideal in up to 40% of the cases because of lack of hands-on experience of residents. To try and solve the issue we propose a medical simulator that will merge haptic feedback with hardware components. Vibro-tactile feedback has been proven useful in medical simulations and could give a more complete and realistic experience to the training surgeon, either as supplementary information to the force feedback or as stand alone information. In order to feed back the vibro-tactile information to the user, the haptic device has to be instrumentalized with appropriate custom-made hardware.
Keywords
Vibro-tactile feedback, Haptic feedback, Medical robotics, Surgical simulators
Labels
Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-02-26 , Earliest start: 2025-03-01
Organization Bio-Inspired RObots for MEDicine-Laboratory (BIROMED-Lab)
Hosts Gerig Nicolas, Dr. , Sommerhalder Michael , Ettori Sara Lisa Margherita
Topics Engineering and Technology
Leveraging Human Motion Data from Videos for Humanoid Robot Motion Learning
The advancement in humanoid robotics has reached a stage where mimicking complex human motions with high accuracy is crucial for tasks ranging from entertainment to human-robot interaction in dynamic environments. Traditional approaches in motion learning, particularly for humanoid robots, rely heavily on motion capture (MoCap) data. However, acquiring large amounts of high-quality MoCap data is both expensive and logistically challenging. In contrast, video footage of human activities, such as sports events or dance performances, is widely available and offers an abundant source of motion data. Building on recent advancements in extracting and utilizing human motion from videos, such as the method proposed in WHAM (refer to the paper "Learning Physically Simulated Tennis Skills from Broadcast Videos"), this project aims to develop a system that extracts human motion from videos and applies it to teach a humanoid robot how to perform similar actions. The primary focus will be on extracting dynamic and expressive motions from videos, such as soccer player celebrations, and using these extracted motions as reference data for reinforcement learning (RL) and imitation learning on a humanoid robot.
Labels
Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-02-25
Applications limited to ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne
Organization ETH Competence Center - ETH AI Center
Hosts Li Chenhao , Kaufmann Manuel , Li Chenhao , Li Chenhao , Kaufmann Manuel , Li Chenhao
Topics Engineering and Technology
Learning Agile Dodgeball Behaviors for Humanoid Robots
Agility and rapid decision-making are vital for humanoid robots to safely and effectively operate in dynamic, unstructured environments. In human contexts—whether in crowded spaces, industrial settings, or collaborative environments—robots must be capable of reacting to fast, unpredictable changes in their surroundings. This includes not only planned navigation around static obstacles but also rapid responses to dynamic threats such as falling objects, sudden human movements, or unexpected collisions. Developing such reactive capabilities in legged robots remains a significant challenge due to the complexity of real-time perception, decision-making under uncertainty, and balance control. Humanoid robots, with their human-like morphology, are uniquely positioned to navigate and interact with human-centered environments. However, achieving fast, dynamic responses—especially while maintaining postural stability—requires advanced control strategies that integrate perception, motion planning, and balance control within tight time constraints. The task of dodging fast-moving objects, such as balls, provides an ideal testbed for studying these capabilities. It encapsulates several core challenges: rapid object detection and trajectory prediction, real-time motion planning, dynamic stability maintenance, and reactive behavior under uncertainty. Moreover, it presents a simplified yet rich framework to investigate more general collision avoidance strategies that could later be extended to complex real-world interactions. In robotics, reactive motion planning for dynamic environments has been widely studied, but primarily in the context of wheeled robots or static obstacle fields. Classical approaches focus on precomputed motion plans or simple reactive strategies, often unsuitable for highly dynamic scenarios where split-second decisions are critical. In the domain of legged robotics, maintaining balance while executing rapid, evasive maneuvers remains a challenging problem. Previous work on dynamic locomotion has addressed agile behaviors like running, jumping, or turning (e.g., Hutter et al., 2016; Kim et al., 2019), but these movements are often planned in advance rather than triggered reactively. More recent efforts have leveraged reinforcement learning (RL) to enable robots to adapt to dynamic environments, demonstrating success in tasks such as obstacle avoidance, perturbation recovery, and agile locomotion (Peng et al., 2017; Hwangbo et al., 2019). However, many of these approaches still struggle with real-time constraints and robustness in high-speed, unpredictable scenarios. Perception-driven control in humanoids, particularly for tasks requiring fast reactions, has seen advances through sensor fusion, visual servoing, and predictive modeling. For example, integrating vision-based object tracking with dynamic motion planning has enabled robots to perform tasks like ball catching or blocking (Ishiguro et al., 2002; Behnke, 2004). Yet, dodging requires a fundamentally different approach: instead of converging toward an object (as in catching), the robot must predict and strategically avoid the object’s trajectory while maintaining balance—often in the presence of limited maneuvering time. Dodgeball-inspired robotics research has been explored in limited contexts, primarily using wheeled robots or simplified agents in simulations. Few studies have addressed the challenges of high-speed evasion combined with the complexities of humanoid balance and multi-joint coordination. This project aims to bridge that gap by developing learning-based methods that enable humanoid robots to reactively avoid fast-approaching objects in real time, while preserving stability and agility.
Labels
Master Thesis
Description
Goal
Contact Details
More information
Published since: 2025-02-25
Applications limited to ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne
Organization ETH Competence Center - ETH AI Center
Hosts Li Chenhao , Li Chenhao , Li Chenhao , Li Chenhao
Topics Engineering and Technology
Learning Real-time Human Motion Tracking on a Humanoid Robot
Humanoid robots, designed to mimic the structure and behavior of humans, have seen significant advancements in kinematics, dynamics, and control systems. Teleoperation of humanoid robots involves complex control strategies to manage bipedal locomotion, balance, and interaction with environments. Research in this area has focused on developing robots that can perform tasks in environments designed for humans, from simple object manipulation to navigating complex terrains. Reinforcement learning has emerged as a powerful method for enabling robots to learn from interactions with their environment, improving their performance over time without explicit programming for every possible scenario. In the context of humanoid robotics and teleoperation, RL can be used to optimize control policies, adapt to new tasks, and improve the efficiency and safety of human-robot interactions. Key challenges include the high dimensionality of the action space, the need for safe exploration, and the transfer of learned skills across different tasks and environments. Integrating human motion tracking with reinforcement learning on humanoid robots represents a cutting-edge area of research. This approach involves using human motion data as input to train RL models, enabling the robot to learn more natural and human-like movements. The goal is to develop systems that can not only replicate human actions in real-time but also adapt and improve their responses over time through learning. Challenges in this area include ensuring real-time performance, dealing with the variability of human motion, and maintaining stability and safety of the humanoid robot.
Keywords
real-time, humanoid, reinforcement learning, representation learning
Labels
Master Thesis
Description
Contact Details
More information
Published since: 2025-02-25
Organization ETH Competence Center - ETH AI Center
Hosts Li Chenhao , Li Chenhao , Li Chenhao , Li Chenhao
Topics Information, Computing and Communication Sciences
Loosely Guided Reinforcement Learning for Humanoid Parkour
Humanoid robots hold the promise of navigating complex, human-centric environments with agility and adaptability. However, training these robots to perform dynamic behaviors such as parkour—jumping, climbing, and traversing obstacles—remains a significant challenge due to the high-dimensional state and action spaces involved. Traditional Reinforcement Learning (RL) struggles in such settings, primarily due to sparse rewards and the extensive exploration needed for complex tasks. This project proposes a novel approach to address these challenges by incorporating loosely guided references into the RL process. Instead of relying solely on task-specific rewards or complex reward shaping, we introduce a simplified reference trajectory that serves as a guide during training. This trajectory, often limited to the robot's base movement, reduces the exploration burden without constraining the policy to strict tracking, allowing the emergence of diverse and adaptable behaviors. Reinforcement Learning has demonstrated remarkable success in training agents for tasks ranging from game playing to robotic manipulation. However, its application to high-dimensional, dynamic tasks like humanoid parkour is hindered by two primary challenges: Exploration Complexity: The vast state-action space of humanoids leads to slow convergence, often requiring millions of training steps. Reward Design: Sparse rewards make it difficult for the agent to discover meaningful behaviors, while dense rewards demand intricate and often brittle design efforts. By introducing a loosely guided reference—a simple trajectory representing the desired flow of the task—we aim to reduce the exploration space while maintaining the flexibility of RL. This approach bridges the gap between pure RL and demonstration-based methods, enabling the learning of complex maneuvers like climbing, jumping, and dynamic obstacle traversal without heavy reliance on reward engineering or exact demonstrations.
Keywords
humanoid, reinforcement learning, loosely guided
Labels
Master Thesis
Description
Contact Details
More information
Published since: 2025-02-25
Organization ETH Competence Center - ETH AI Center
Hosts Li Chenhao , Li Chenhao , Li Chenhao , Li Chenhao
Topics Information, Computing and Communication Sciences
Learning World Models for Legged Locomotion
Model-based reinforcement learning learns a world model from which an optimal control policy can be extracted. Understanding and predicting the forward dynamics of legged systems is crucial for effective control and planning. Forward dynamics involve predicting the next state of the robot given its current state and the applied actions. While traditional physics-based models can provide a baseline understanding, they often struggle with the complexities and non-linearities inherent in real-world scenarios, particularly due to the varying contact patterns of the robot's feet with the ground. The project aims to develop and evaluate neural network-based models for predicting the dynamics of legged environments, focusing on accounting for varying contact patterns and non-linearities. This involves collecting and preprocessing data from various simulation environment experiments, designing neural network architectures that incorporate necessary structures, and exploring hybrid models that combine physics-based predictions with neural network corrections. The models will be trained and evaluated on prediction autoregressive accuracy, with an emphasis on robustness and generalization capabilities across different noise perturbations. By the end of the project, the goal is to achieve an accurate, robust, and generalizable predictive model for the forward dynamics of legged systems.
Keywords
forward dynamics, non-smooth dynamics, neural networks, model-based reinforcement learning
Labels
Master Thesis
Description
Contact Details
More information
Published since: 2025-02-25
Organization Robotic Systems Lab
Hosts Li Chenhao , Li Chenhao , Li Chenhao , Li Chenhao
Topics Engineering and Technology
Exploring upper limb impairments using explainable AI on Virtual Peg Insertion Test data
This thesis aims to apply explainable AI techniques to analyze time series data from the Virtual Peg Insertion Test (VPIT), uncovering additional metrics that describe upper limb impairments in neurological subjects, such as those with stroke, Parkinson's disease, and multiple sclerosis. By preserving the full dimensionality of the data, the project will identify new patterns and insights to aid in understanding motor dysfunctions and support rehabilitation.
Keywords
Machine learning, rehabilitation, neurology, upper limb, impairment, explainable AI, SHAP, novel technology, assessment, computer vision, artificial intelligence
Labels
Master Thesis
Description
Goal
Tasks
Your Profile
Contact Details
More information
Published since: 2025-02-18 , Earliest start: 2025-03-09
Organization Rehabilitation Engineering Lab
Hosts Domnik Nadine
Topics Medical and Health Sciences , Information, Computing and Communication Sciences , Engineering and Technology
Comparing the Virtual Peg Insertion Test (VPIT) with the haptic device Inverse3 for assessing upper limb function
This thesis will compare the Virtual Peg Insertion Test (VPIT) with the Inverse3 haptic device by Haply to evaluate its effectiveness as a tool for assessing upper limb function. The focus will be on comparing both the hardware features and software capabilities to determine if the Inverse3 can serve as a valid alternative to VPIT for clinical assessments.
Keywords
Haptic device, virtual environment, rehabilitation, programming, health technology, assessment, software, hardware
Labels
Collaboration , Master Thesis
Description
Goal
Tasks
Your Profile
Contact Details
More information
Published since: 2025-02-18 , Earliest start: 2025-03-16
Organization Rehabilitation Engineering Lab
Hosts Domnik Nadine
Topics Medical and Health Sciences , Information, Computing and Communication Sciences , Engineering and Technology
Embedded algorithms of IMUs in a neurorehabilitation device
The goal of this project is to help develop embedded firmware for a imu based rehabilitation device. This project is part of the SmartVNS project which utilizes movement-gated control of vagus nerve stimulation for stroke rehabilitation.
Keywords
electrical engineering PCB Embedded systems neurorehabilitation
Labels
Semester Project , Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-02-14 , Earliest start: 2024-01-06 , Latest end: 2024-12-31
Organization Rehabilitation Engineering Lab
Hosts Donegan Dane , Viskaitis Paulius
Topics Medical and Health Sciences , Engineering and Technology
Analysis and modelling of Neurophysiological Data from Multisensory Recordings during aVNS Experiments
Join our research project focused on analysing complex neurophysiological data collected during non-invasive brain stimulation experiments. This project aims to optimise brain stimulation protocols for future stroke rehabilitation by investigating neural responses to various stimulation parameters. The data includes electrocardiogram (ECG), electroencephalogram (EEG), photoplethysmography (PPG), inertial measurement unit (IMU) readings, pupilometry, and galvanic skin response (GSR). We aim to model brain states based on these measurements to define brain circuitry outcomes from stimulation and movement interactions, using advanced techniques like connectivity-based biomarkers. This modeling will help generalise findings to broader brain states, such as valence, attention, and stress.
Labels
Master Thesis
Goal
Tasks
Your Profile
Contact Details
More information
Published since: 2025-02-14 , Earliest start: 2024-08-18 , Latest end: 2025-04-30
Organization Rehabilitation Engineering Lab
Hosts Viskaitis Paulius
Topics Medical and Health Sciences , Mathematical Sciences , Information, Computing and Communication Sciences
Feasibility of RehabCoach on Adherence to Unsupervised Robot-Assisted Therapy – Implementation and Evaluation of Smart Reminders
Adherence to rehabilitation therapy is essential for the recovery of hand functionality in stroke and traumatic brain injury (TBI) patients. However, maintaining engagement outside clinical settings remains a challenge. This project involves a feasibility study to evaluate the adherence of patients using the RehabCoach app as part of a one-week unsupervised robot-assisted program simulation. The study assesses user engagement, app interaction patterns, and the effectiveness of push notifications/smart reminders in sustaining adherence to the training program. The key components of this research are the development of a smart algorithm for triggering push notifications based on specific user behaviors, such as therapy completion or inactivity, to optimize adherence, as well as conducting the study with a few participants.
Keywords
Stroke, Traumatic Brain Injury, Rehabilitation Therapy, Adherence, Push Notifications, Mobile Health App, Interdisciplinary Research, Python, Django
Labels
Semester Project , Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-02-12 , Earliest start: 2025-02-16 , Latest end: 2025-11-30
Organization Rehabilitation Engineering Lab
Hosts Retevoi Alexandra
Topics Medical and Health Sciences , Information, Computing and Communication Sciences , Engineering and Technology
Benefits and challenges of promoting minimally supervised therapy in a rehabilitation clinic
Increasing the therapy time can benefit patients in multiple ways. Group therapy (e.g., technology-assisted) allows clinics to increase the therapy dose for patients without increasing the workload for therapists. However, in practical implementation, some challenges often arise (e.g., patients not liking it) that limit the efficacy of group therapy. The aim of this project is to gain practical experience with group therapy sessions (i.e., supporting group therapy sessions at the Clinica Hildebrand in Brissago), identify barriers and benefits related to group therapy, and propose guidelines to improve the integration of group therapy in clinical practice.
Keywords
practical internship, research internship, clinical work, technology-assisted therapy, group therapy, minimally supervised therapy
Labels
Internship
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-02-12 , Earliest start: 2025-03-01 , Latest end: 2025-10-31
Applications limited to ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne , Università della Svizzera italiana
Organization Rehabilitation Engineering Lab
Hosts Devittori Giada
Topics Medical and Health Sciences
Supervised learning for loco-manipulation
To spot arm operations, we propose a multi-phase approach combining supervised learning and reinforcement learning (RL). First, we will employ supervised learning to develop a model for solving inverse kinematics (IK), enabling precise joint angle calculations from desired end-effector pose. Next, we will utilize another supervised learning technique to build a collision avoidance model, trained to predict and avoid self-collisions based on arm configurations and environmental data. With these pre-trained networks, we will then integrate RL to generate dynamic and safe arm-motion plans. The RL agent will leverage the IK and collision avoidance models to optimize arm trajectories, ensuring efficient and collision-free movements. This entire pipeline could be back propagated while promising to enhance the accuracy, safety, and flexibility of robotic arm operations in complex environments.
Keywords
Spot, Supervised learning, loco-manipulation
Labels
Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-02-10 , Earliest start: 2025-02-10 , Latest end: 2026-03-01
Organization Robotic Systems Lab
Hosts Mirrazavi Sina
Topics Information, Computing and Communication Sciences
Model-Based Reinforcement Learning for Loco-manipulation
This project aims to develop a model-based reinforcement learning (RL) framework to enable quadruped robots to perform dynamic locomotion and manipulation simultaneously by leveraging advanced model-based RL algorithms such as DeamerV3, TDMPC2 and SAM-RL. We will develop control policies that can predict future states and rewards, enabling the robot to adapt its behavior on-the-fly. The primary focus will be on achieving stable and adaptive walking patterns while reaching and grasping objects. The outcome will provide insights into the integration of complex behaviors in robotic systems, with potential applications in service robotics and automated object handling.
Labels
Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-02-10 , Earliest start: 2025-02-10 , Latest end: 2026-02-10
Organization Robotic Systems Lab
Hosts Mirrazavi Sina
Topics Information, Computing and Communication Sciences
Integrating OpenVLA for Vision-Language-Driven Loco-Manipulation robotics scenarios
This thesis proposes to integrate and adapt the OpenVLA (Open-Source Vision-Language-Action) model to control the Spot robotic arm for performing complex grasping and placing tasks. The study will focus on enabling the robot to recognize, grasp, and organize various toy-sized kitchen items based on human instructions. By leveraging OpenVLA's robust multimodal capabilities, this project aims to bridge the gap between human intent and robotic actions, enabling seamless task execution in unstructured environments. The research will explore the feasibility of fine-tuning OpenVLA for task-specific operations and evaluate its performance in real-world scenarios, providing valuable insights for advancing multimodal robotics.
Labels
Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-02-10 , Earliest start: 2025-02-10 , Latest end: 2026-02-10
Organization Robotic Systems Lab
Hosts Mirrazavi Sina
Topics Information, Computing and Communication Sciences
Differentiable Simulation for Precise End-Effector Tracking
Unlock the potential of differentiable simulation on ALMA, a quadrupedal robot equipped with a robotic arm. Differentiable simulation enables precise gradient-based optimization, promising greater tracking accuracy and efficiency compared to standard reinforcement learning approaches. This project dives into advanced simulation and control techniques, paving the way for improvements in robotic trajectory tracking.
Keywords
Differentiable Simulation, Learning, ALMA
Labels
Semester Project , Bachelor Thesis , Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-02-07 , Earliest start: 2025-01-27
Organization Robotic Systems Lab
Hosts Mittal Mayank , Schwarke Clemens , Klemm Victor
Topics Information, Computing and Communication Sciences
Research Assistant in Biosensing for Robotics Care and Body Simulation (~12 months)
Join a team of scientists improving the long-term prognosis and treatment of Spinal Cord Injury (SCI) through mobile and wearable systems and personalized health monitoring. Joining the SCAI Lab part of the Sensory-Motor Systems Lab at ETH, you will have the unique opportunity of working at one of the largest and most prestigious health providers in Switzerland: Swiss Paraplegic Center (SPZ) in Nottwil (LU).
Keywords
App development, Machine Learning, Data bases, Data engineering, Systems Engineering, Data Modelling
Labels
Internship , Lab Practice , Student Assistant / HiWi , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2025-02-06 , Earliest start: 2025-03-03 , Latest end: 2026-12-31
Applications limited to ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne , IBM Research Zurich Lab , Institute for Research in Biomedicine , Hochschulmedizin Zürich , Swiss Institute of Bioinformatics , University of Lucerne , University of Zurich , Zurich University of Applied Sciences , Zurich University of the Arts , Lucerne University of Applied Sciences and Arts , Berner Fachhochschule
Organization Spinal Cord Injury & Artificial Intelligence Lab
Hosts Paez Diego, Dr. , Paez Diego, Dr. , Paez Diego, Dr.
Topics Information, Computing and Communication Sciences
Master's Thesis: AI-powered nap detection from Fitbit data
The uprise of consumer-grade fitness trackers has opened the doors to long-term activity monitoring in the wild in research and clinics. However, Fitbit does not identify napping episodes shorter than 90 minutes. Hence, there is a need to establish a robust algorithm to detect naps.
Keywords
Data analysis, machine learning, signal processing, wearables, Fitbit, naps detection
Labels
Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2025-02-06 , Earliest start: 2025-02-10 , Latest end: 2025-08-31
Organization Spinal Cord Injury & Artificial Intelligence Lab
Hosts Gnarra Oriella , Gnarra Oriella
Topics Information, Computing and Communication Sciences , Engineering and Technology
Modeling and Simulation for Earthwork in Digital Twin
In this work, we aim to build a digital twin of our autonomous hydraulic excavator, leveraging Mathworks technology for high-fidelity modeling. This will be used in the future to test and benchmark our learning-based controllers.
Keywords
Modeling, Hydraulics, Excavation, Industry
Labels
Semester Project , Master Thesis , ETH Zurich (ETHZ)
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-02-06 , Earliest start: 2025-03-03
Organization Robotic Systems Lab
Hosts Spinelli Filippo , Nan Fang
Topics Information, Computing and Communication Sciences , Engineering and Technology
Reinforcement Learning for Excavation Planning In Terra
We aim to develop a reinforcement learning-based global excavation planner that can plan for the long term and execute a wide range of excavation geometries. The system will be deployed on our legged excavator.
Keywords
Keywords: Reinforcement learning, task planning
Labels
Semester Project , Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-02-03 , Earliest start: 2025-04-01 , Latest end: 2025-08-31
Organization Robotic Systems Lab
Hosts Terenzi Lorenzo
Topics Information, Computing and Communication Sciences
Model Based Reinforcement Learning
We want to train an excavator agent to learn in a variety of soil using a fast, GPU-accelerated soil particle simulator in Isaac Sim.
Keywords
particle simulation, omniverse, warp, reinforcement learning, model based reinforcement learning.
Labels
Semester Project , Master Thesis
Description
Work Packages
Requirements
More information
Published since: 2025-02-03 , Earliest start: 2025-02-28 , Latest end: 2025-08-31
Organization Robotic Systems Lab
Hosts Egli Pascal Arturo , Terenzi Lorenzo
Topics Information, Computing and Communication Sciences , Engineering and Technology
Reinforcement Learning for Particle-Based Excavation in Isaac Sim
We want to train RL agents on our new particle simulator, accelerated on the GPU via warp in Isaac sim.
Keywords
particle simulation, omniverse, warp, reinforcement learning
Labels
Semester Project , Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-02-03 , Earliest start: 2025-04-01 , Latest end: 2025-09-30
Organization Robotic Systems Lab
Hosts Egli Pascal Arturo , Mittal Mayank , Terenzi Lorenzo
Topics Information, Computing and Communication Sciences
Perceptive Reinforcement Learning for Exavation
In this project, our goal is to leverage precomputed embeddings(VAE in Isaacsim) from 3D earthworks scene reconstructions to train reinforcement learning agents. These embeddings, derived from incomplete point cloud data and reconstructed using an encoder-decoder neural network, will serve as latent representations. The main emphasis is on utilizing these representations to develop and train reinforcement learning policies for digging tasks.
Keywords
LIDAR, 3D reconstruction, Isaac gym, deep learning, perception, reinforcement learning
Labels
Semester Project , Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-02-03 , Earliest start: 2025-04-01 , Latest end: 2025-08-31
Organization Robotic Systems Lab
Hosts Höller David , Terenzi Lorenzo
Topics Information, Computing and Communication Sciences
Depth Estimation with Event-Based Cameras on nano-UAVs
develop a neural network for depth estimation on nano-drones, using miniaturized event-based cameras, for autonomous navigation
Keywords
nano-drones, robotics, autonomous navigation, depth-estimation, event-based cameras, neuromorphic, neuromorphic vision, embedded systems, tinyML, artificial intelligence, deep learning.
Labels
Semester Project , Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2025-02-03 , Earliest start: 2025-02-03 , Latest end: 2025-12-31
Organization Digital Circuits and Systems (Benini)
Hosts Lamberti Lorenzo
Topics Engineering and Technology
Digital HW design: accelerator for Event-Based Convolutional Neural Network
Digital HW design: accelerator for Event-Based Convolutional Neural Network
Keywords
Digital HW design, PULP, RISC-V
Labels
Semester Project , Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
Description
Contact Details
More information
Published since: 2025-02-03 , Earliest start: 2025-02-03 , Latest end: 2025-07-31
Organization Digital Circuits and Systems (Benini)
Hosts Lamberti Lorenzo
Topics Engineering and Technology
Artificial intelligence and perception on insect sized robots
Maze escape with tiny AI-based autonomous bugs
Keywords
Insect-sized robots, robotics, artificial intelligence, embedded systems, autonomous navigation
Labels
Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
Description
Contact Details
More information
Published since: 2025-02-03 , Latest end: 2025-07-01
Organization Digital Circuits and Systems (Benini)
Hosts Lamberti Lorenzo
Topics Engineering and Technology
Autonomous nano-drone racing CNN on next-gen ULP SoC
Develop a visual-based neural network for autonomous nano-drone racing
Keywords
nano-drones, robotics, autonomous navigation, embedded systems,artificial intellicence, deep learning, microcontroller
Labels
Semester Project , Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
Description
Contact Details
More information
Published since: 2025-02-03 , Latest end: 2025-07-31
Organization Digital Circuits and Systems (Benini)
Hosts Lamberti Lorenzo
Topics Engineering and Technology
Reiforcement Learning of Pretrained Trasformer Models
We want to train RL agents on our new particle simulator, accelerated on the GPU via warp in Isaac sim.
Keywords
Keywords: particle simulation, omniverse, warp, reinforcement learning
Labels
Semester Project , Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-02-03 , Earliest start: 2025-04-01 , Latest end: 2025-08-31
Organization Robotic Systems Lab
Hosts Terenzi Lorenzo
Topics Information, Computing and Communication Sciences
Multiagent Reinforcement Learning in Terra
We want to train multiple agents in the Terra environment, a fully end-to-end GPU-accelerated environment for RL training.
Keywords
multiagent reinforcement learning, jax, deep learning, planning
Labels
Semester Project , Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-02-03 , Earliest start: 2025-04-01 , Latest end: 2025-09-16
Organization Robotic Systems Lab
Hosts Terenzi Lorenzo
Topics Information, Computing and Communication Sciences
Hand-object 3D reconstruction from Internet videos (Computer Vision)
The goal of the project is to reconstruct 3D hands and objects from internet videos.
Keywords
Computer vision, VR/AR, 3D reconstruction, 3D pose estimation, machine learning, neural networks, human-object interactions
Labels
Semester Project , Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-02-01 , Earliest start: 2024-11-18 , Latest end: 2025-08-20
Applications limited to ETH Zurich , Department of Computer Science , Eberhard Karls Universität Tübingen
Organization Advanced Interactive Technologies
Hosts Fan Zicong
Topics Information, Computing and Communication Sciences
A Bayesian sensor fusion and machine learning approach for robust hand gesture decoding with application to stroke rehabilitation.
About the project: This thesis aims to design a framework for robust fine-motor action decoding using multi-modal (sEMG and depth sensing camera) Bayesian sensor fusion and machine learning approach
Keywords
Bayesian inference, sEMG, depth sensing camera, rehabilitation, machine learning, deep transfer learning
Labels
Semester Project , Bachelor Thesis , Master Thesis
Project Background
Your Task
Your Benefits
Your Profile
Contact Details
More information
Published since: 2025-02-01 , Earliest start: 2025-03-01 , Latest end: 2025-08-01
Organization Sensory-Motor Systems Lab
Hosts Dash Adyasha
Topics Medical and Health Sciences , Information, Computing and Communication Sciences , Engineering and Technology
Part-time Job at a MedTech Startup: Implementation of Wheelchair Durability Testing Infrastructure
Versive, an ETH Spinoff in formation, is developing innovative manual wheelchairs using their disruptive "steering-by-leaning" principle (as seen here: https://www.youtube.com/watch?v=HrJFH3MLlaw ). Currently, we're planning a market launch at the end of 2026 or beginning of 2027. As part of an iterative design process, we need to be able to run essential durability tests ensuring a quick CE marking (medical device class 1) procedure later on. Besides static stability, flammability and biocompatibility, the ISO tests require mechanical durability testing. These include a rolling test (several days on a roller, under load) as well as a drop test (6666 drops from 15cm, fully loaded). For this, we are looking to implement our own testing infrastructure.
Keywords
Wheelchair, Steering-by-Leaning, Testing, Test Bench, ISO Testing, CE Mark, MedTech, Assistive Technology, Startup
Labels
Collaboration , Internship , Lab Practice , Student Assistant / HiWi
Description
Goal
Contact Details
More information
Published since: 2025-01-29 , Earliest start: 2025-02-01 , Latest end: 2025-06-30
Organization Taylor Group / Laboratory for Movement Biomechanics
Hosts Togni Reto
Topics Medical and Health Sciences , Engineering and Technology
Propose Your Own Robotics Challenge
This project invites you to step into the role of an innovator, encouraging you to identify challenges you are passionate about within the field of robotics. Rather than working on predefined problems, you will have the freedom to propose your own project ideas, address real-world issues, or explore cutting-edge topics. This project allows you to define your own research journey.
Keywords
Robotics, Research
Labels
Semester Project , Bachelor Thesis , Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-01-28 , Earliest start: 2025-01-27
Organization Robotic Systems Lab
Hosts Schwarke Clemens , Bjelonic Filip , Klemm Victor
Topics Information, Computing and Communication Sciences
Data Driven Simulation for End-to-End Navigation
Investigate how neural rendering can become the backbone of comprehensive, next generation data-driven simulation
Keywords
Neural rendering, Simulation
Labels
Internship , Master Thesis
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2025-01-24 , Earliest start: 2025-01-27
Organization Robotic Systems Lab
Hosts Kneip Laurent
Topics Information, Computing and Communication Sciences , Engineering and Technology
Master Thesis: Data Analysis of Wearable and Nearable Sensors Data for Classification of Activities of Daily Living
This project aims to develop a novel algorithm for tracking a person's health condition changes using daily life wearable sensor data, biosignals, and information from nearable sensors. With the Life-long-logging system, we want to provide meaningful data for medical staff and directly engage patients and their caregivers.
Keywords
Data analysis, Machine Learning, Wearable Sensors
Labels
Semester Project , Internship , Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2025-01-22 , Earliest start: 2025-02-03 , Latest end: 2025-09-30
Organization Spinal Cord Injury & Artificial Intelligence Lab
Hosts Gnarra Oriella , Gnarra Oriella
Topics Information, Computing and Communication Sciences , Engineering and Technology
Master Thesis: Data Analysis of Wearable and Nearable Sensors Data within the StrongAge Cohort Study
The StrongAge Dataset, collected over one year, provides a rich data repository from unobtrusive, contactless technologies combined with validated mood and cognition questionnaires. This project aims to uncover digital biomarkers that can transform elderly care, addressing critical research questions related to sleep, cognition, physical activity, and environmental influences.
Keywords
Data analysis, Machine learning, Wearable and Nearable Sensors Data
Labels
Semester Project , Internship , Lab Practice , Bachelor Thesis , Master Thesis , Student Assistant / HiWi , ETH Zurich (ETHZ)
Description
Goal
Contact Details
More information
Published since: 2025-01-22 , Earliest start: 2025-02-03 , Latest end: 2025-09-30
Organization Spinal Cord Injury & Artificial Intelligence Lab
Hosts Gnarra Oriella , Gnarra Oriella
Topics Information, Computing and Communication Sciences , Engineering and Technology
Reconstructing liquids from multiple views with 3D Gaussian Splatting
This project reconstructs liquids from multi-view imagery, segmenting fluid regions using methods like Mask2Former and reconstructing static scenes with 3D Gaussian Splatting or Mast3r. The identified fluid clusters initialize a particle-based simulation, refined for temporal consistency and enhanced by optional thermal data and visual language models for fluid properties.
Keywords
3D reconstruction, Gaussian Splatting, physics simulation
Labels
Semester Project , Master Thesis
Description
More information
Published since: 2025-01-08
Organization Computer Vision and Geometry Group
Hosts Barath Daniel , Tsalicoglou Christina
Topics Information, Computing and Communication Sciences
How to Touch: Exploring Tactile Representations for Reinforcement Learning
Developing and benchmarking tactile representations for dexterous manipulation tasks using reinforcement learning.
Keywords
Reinforcement Learning, Dexterous Manipulation, Tactile Sensing
Labels
Semester Project , Bachelor Thesis , Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2025-01-08 , Earliest start: 2024-12-15 , Latest end: 2025-06-01
Applications limited to ETH Zurich
Organization Robotic Systems Lab
Hosts Bhardwaj Arjun , Zurbrügg René
Topics Information, Computing and Communication Sciences
A generalized sEMG-based gesture recognition framework using deep learning approach
This thesis aims to develop a generalizable (user-invariant and session-invariant) gesture recognition framework using deep neural networks
Keywords
sEMG, deep learning, manifold, gesture decoding
Labels
Master Thesis
Project Background
Your Task
Your Benefits
Your Profile
Contact Details
More information
Published since: 2025-01-05 , Earliest start: 2025-02-01 , Latest end: 2026-03-01
Organization Sensory-Motor Systems Lab
Hosts Dash Adyasha
Topics Information, Computing and Communication Sciences , Engineering and Technology
VR-based Motor-cognitive rehabilitation with multi-modal sensing
This thesis aims to design a pool of virtual-reality based rehabilitation tasks for simultaneous motor-cognitive rehabilitation for stroke patients
Keywords
virtual reality, sEMG, depth sensing camera, rehabilitation,
Labels
Semester Project , Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
Project Background
Your Task
Your Benefits
Your Profile
Contact Details
More information
Published since: 2025-01-02 , Earliest start: 2025-03-01 , Latest end: 2025-12-31
Organization Sensory-Motor Systems Lab
Hosts Dash Adyasha
Topics Medical and Health Sciences , Information, Computing and Communication Sciences , Engineering and Technology , Behavioural and Cognitive Sciences
Unraveling Calcium Dynamics and Immune Interactions in Bone Graft Substitute Environments through Advanced Ratiometric Imaging
This project endeavors to explore the dynamic interplay among calcium ions, bone graft substitutes, and resident immune cells in both orthotopic and ectopic environments, employing advanced ratiometric imaging techniques.
Keywords
Bone Graft Substitute, Calcium, Ratiometric Imaging, Immune Cells, in vitro, in vivo, Intravital Microscopy
Labels
Semester Project , Internship , Bachelor Thesis , Master Thesis
Description
Goal
Contact Details
More information
Published since: 2024-12-24 , Earliest start: 2024-10-01 , Latest end: 2025-06-30
Organization Müller Group / Laboratory for Bone Biomechanics
Hosts Wissmann Stefanie
Topics Engineering and Technology , Biology
Developing a real-time state-machine for closed-loop brain-machine interfaces
Developing a state-machine Simulink model to be deployed at MathWorks SpeedGoat real-time target machine for closed-loop brain-machine interface (BMI). The state-machine will control the closed-loop BMI peripherals and synchronise the data flow. Peripherals include neural recorders & stimulators, data analysis cluster, video cameras and experimental chamber. Experimental chamber (variety of servos, steppers, sensors etc.) will be controlled with built-in FPGA and GPIO of SpeedGoat machine. Other peripherals are connected with serial bus. Acquired data needs to be organized and stored in datasink unit. Skills: Matlab Simulink, state-machines, FPGA programming, serial communication protocols, data synchronisation Please send your CV and transcript along with your application.
Keywords
Brain-machine interfaces, neural recording, neural stimulation, state-machines, real-time systems, serial communication, FPGA
Labels
Semester Project , Internship , Bachelor Thesis , Master Thesis
Contact Details
More information
Published since: 2024-12-20
Applications limited to EPFL - Ecole Polytechnique Fédérale de Lausanne , ETH Zurich , University of Zurich , Wyss Translational Center Zurich , Paul Scherrer Institute , CERN , Empa
Organization Neurotechnology
Hosts Özil Eminhan
Topics Engineering and Technology
Developing a GUI for cross-modality brain-machine interfaces
Programming a graphical user interface (e.g. in Qt/C++) which can handle and process the data acquired in our brain-machine interface (BMI) experiments. The data includes high-density brain activity recordings from hundreds of recording channels, neural-stimulation events, 3D&4D data coming from MRI scans of the subject implanted with BMI. The backend will be programmed in Python where you also need to connect supporting tools (e.g. Blender) via Python. Please send an email with your CV and transcript of records attached.
Keywords
Brain-machine interface, BMI, graphical user interface, GUI, Qt, C, C++, Python, Blender, MRI, electrophysiology
Labels
Semester Project , Internship , Bachelor Thesis , Master Thesis , ETH Zurich (ETHZ)
Contact Details
More information
Published since: 2024-12-19
Applications limited to Empa , EPFL - Ecole Polytechnique Fédérale de Lausanne , ETH Zurich , IBM Research Zurich Lab , Paul Scherrer Institute , University of Zurich , Wyss Translational Center Zurich
Organization Neurotechnology
Hosts Özil Eminhan
Topics Information, Computing and Communication Sciences , Engineering and Technology
Digital Twin for Spot's Home
MOTIVATION ⇾ Creating a digital twin of the robot's environment is crucial for several reasons: 1. Simulate Different Robots: Test various robots in a virtual environment, saving time and resources. 2. Accurate Evaluation: Precisely assess robot interactions and performance. 3. Enhanced Flexibility: Easily modify scenarios to develop robust systems. 4. Cost Efficiency: Reduce costs by identifying issues in virtual simulations. 5. Scalability: Replicate multiple environments for comprehensive testing. PROPOSAL We propose to create a digital twin of our Semantic environment, designed in your preferred graphics Platform to be able to simulate Reinforcement Learning agents in the digital environment, to create a unified evaluation platform for robotic tasks.
Keywords
Digital Twin, Robotics
Labels
Semester Project , Master Thesis
Contact Details
More information
Published since: 2024-12-17 , Earliest start: 2025-01-05
Applications limited to University of Zurich , ETH Zurich , EPFL - Ecole Polytechnique Fédérale de Lausanne
Organization Computer Vision and Geometry Group
Hosts Blum Hermann , Portela Tifanny , Bauer Zuria, Dr. , Trisovic Jelena
Topics Information, Computing and Communication Sciences
KALLAX Benchmark: Evaluating Household Tasks
Motivation ⇾ There are three ways to evaluate robots for pick-and-place tasks at home: 1. Simulation setups: High reproducibility but hard to simulate real-world complexities and perception noise. 2. Competitions: Good for comparing overall systems but require significant effort and can't be done frequently. 3. Custom lab setups: Common but lead to overfitting and lack comparability between labs. Proposal ⇾ We propose using IKEA furniture to create standardized, randomized setups that researchers can easily replicate. E.g, a 4x4 KALLAX unit with varying door knobs and drawer positions, generating tasks like "move the cup from the upper right shelf into the black drawer." This prevents overfitting and allows for consistent evaluation across different labs.
Keywords
Benchmakr, Robotics, pick-and-place
Labels
Semester Project , Master Thesis
Contact Details
More information
Published since: 2024-12-17 , Earliest start: 2025-01-06
Applications limited to University of Zurich , ETH Zurich , Swiss National Science Foundation , EPFL - Ecole Polytechnique Fédérale de Lausanne
Organization Computer Vision and Geometry Group
Hosts Blum Hermann , Bauer Zuria, Dr. , Zurbrügg René
Topics Information, Computing and Communication Sciences
Flexible Wireless Sensing Node for Continuous Body Monitoring
Advancements in sensor technology, low-power mixed-signal/RF circuits, and Wireless Sensor Networks (WSNs) have enabled the creation of compact, cost-effective solutions for healthcare applications. A notable development in this field is the Body Sensor Network, which is designed to monitor the human body for healthcare purposes.
Keywords
Embedded systems, PCB design, Firmware development, Data Analysis
Labels
Semester Project , Internship , Bachelor Thesis
Description
Goal
Contact Details
More information
Published since: 2024-12-14 , Earliest start: 2025-01-12
Organization Digital Circuits and Systems (Benini)
Hosts Spacone Giusy
Topics Engineering and Technology
Visual Language Models for Long-Term Planning
This project uses Visual Language Models (VLMs) for high-level planning and supervision in construction tasks, enabling task prioritization, dynamic adaptation, and multi-robot collaboration for excavation and site management. prioritization, dynamic adaptation, and multi-robot collaboration for excavation and site management
Keywords
Visual Language Models, Long-term planning, Robotics
Labels
Semester Project , Master Thesis
Description
Work Packages
Contact Details
More information
Published since: 2024-12-06 , Earliest start: 2025-03-31 , Latest end: 2025-10-29
Organization Robotic Systems Lab
Hosts Terenzi Lorenzo
Topics Information, Computing and Communication Sciences
Diffusion-based Shared Autonomy System for Telemanipulation
Robots may not be able to complete tasks fully autonomously in unstructured or unseen environments, however direct teleoperation from human operators may also be challenging due to the difficulty of providing full situational awareness to the operator as well as degradation in communication leading to the loss of control authority. This motivates the use of shared autonomy for assisting the operator thereby enhancing the performance during the task. In this project, we aim to develop a shared autonomy framework for teleoperation of manipulator arms, to assist non-expert users or in the presence of degraded communication. Imitation learning, such as diffusion models, have emerged as a popular and scalable approach for learning manipulation tasks [1, 2]. Additionally, recent works have combined this with partial diffusion to enable shared autonomy [3]. However, the tasks were restricted to simple 2D domains. In this project, we wish to extend previous work in the lab using diffusion-based imitation learning, to enable shared autonomy for non-expert users to complete unseen tasks or in degraded communication environments.
Keywords
Imitation learning, Robotics, Manipulation, Teleoperation
Labels
Semester Project , ETH Zurich (ETHZ)
Description
Work Packages
Requirements
Contact Details
More information
Published since: 2024-12-02 , Earliest start: 2024-11-01 , Latest end: 2025-11-01
Applications limited to ETH Zurich , University of Zurich
Organization Robotic Systems Lab
Hosts Elanjimattathil Aravind
Topics Information, Computing and Communication Sciences , Engineering and Technology
Lifelike Agility on ANYmal by Learning from Animals
The remarkable agility of animals, characterized by their rapid, fluid movements and precise interaction with their environment, serves as an inspiration for advancements in legged robotics. Recent progress in the field has underscored the potential of learning-based methods for robot control. These methods streamline the development process by optimizing control mechanisms directly from sensory inputs to actuator outputs, often employing deep reinforcement learning (RL) algorithms. By training in simulated environments, these algorithms can develop locomotion skills that are subsequently transferred to physical robots. Although this approach has led to significant achievements in achieving robust locomotion, mimicking the wide range of agile capabilities observed in animals remains a significant challenge. Traditionally, manually crafted controllers have succeeded in replicating complex behaviors, but their development is labor-intensive and demands a high level of expertise in each specific skill. Reinforcement learning offers a promising alternative by potentially reducing the manual labor involved in controller development. However, crafting learning objectives that lead to the desired behaviors in robots also requires considerable expertise, specific to each skill.
Keywords
learning from demonstrations, imitation learning, reinforcement learning
Labels
Master Thesis
Description
Contact Details
More information
Published since: 2024-11-26
Organization ETH Competence Center - ETH AI Center
Hosts Li Chenhao , Li Chenhao , Klemm Victor
Topics Information, Computing and Communication Sciences
Pushing the Limit of Quadruped Running Speed with Autonomous Curriculum Learning
The project aims to explore curriculum learning techniques to push the limits of quadruped running speed using reinforcement learning. By systematically designing and implementing curricula that guide the learning process, the project seeks to develop a quadruped controller capable of achieving the fastest possible forward locomotion. This involves not only optimizing the learning process but also ensuring the robustness and adaptability of the learned policies across various running conditions.
Keywords
curriculum learning, fast locomotion
Labels
Master Thesis
Description
Contact Details
More information
Published since: 2024-11-26
Organization Robotic Systems Lab
Hosts Li Chenhao , Bagatella Marco , Li Chenhao , Li Chenhao , Li Chenhao
Topics Engineering and Technology
Humanoid Locomotion Learning and Finetuning from Human Feedback
In the burgeoning field of deep reinforcement learning (RL), agents autonomously develop complex behaviors through a process of trial and error. Yet, the application of RL across various domains faces notable hurdles, particularly in devising appropriate reward functions. Traditional approaches often resort to sparse rewards for simplicity, though these prove inadequate for training efficient agents. Consequently, real-world applications may necessitate elaborate setups, such as employing accelerometers for door interaction detection, thermal imaging for action recognition, or motion capture systems for precise object tracking. Despite these advanced solutions, crafting an ideal reward function remains challenging due to the propensity of RL algorithms to exploit the reward system in unforeseen ways. Agents might fulfill objectives in unexpected manners, highlighting the complexity of encoding desired behaviors, like adherence to social norms, into a reward function. An alternative strategy, imitation learning, circumvents the intricacies of reward engineering by having the agent learn through the emulation of expert behavior. However, acquiring a sufficient number of high-quality demonstrations for this purpose is often impractically costly. Humans, in contrast, learn with remarkable autonomy, benefiting from intermittent guidance from educators who provide tailored feedback based on the learner's progress. This interactive learning model holds promise for artificial agents, offering a customized learning trajectory that mitigates reward exploitation without extensive reward function engineering. The challenge lies in ensuring the feedback process is both manageable for humans and rich enough to be effective. Despite its potential, the implementation of human-in-the-loop (HiL) RL remains limited in practice. Our research endeavors to significantly lessen the human labor involved in HiL learning, leveraging both unsupervised pre-training and preference-based learning to enhance agent development with minimal human intervention.
Keywords
reinforcement learning from human feedback, preference learning
Labels
Master Thesis
Description
Goal
Contact Details
More information
Published since: 2024-11-26
Organization ETH Competence Center - ETH AI Center
Hosts Li Chenhao , Li Chenhao , Chen Xin , Li Chenhao
Topics Information, Computing and Communication Sciences , Engineering and Technology
Online Safe Locomotion Learning in the Wild
Reinforcement learning (RL) can potentially solve complex problems in a purely data-driven manner. Still, the state-of-the-art in applying RL in robotics, relies heavily on high-fidelity simulators. While learning in simulation allows to circumvent sample complexity challenges that are common in model-free RL, even slight distribution shift ("sim-to-real gap") between simulation and the real system can cause these algorithms to easily fail. Recent advances in model-based reinforcement learning have led to superior sample efficiency, enabling online learning without a simulator. Nonetheless, learning online cannot cause any damage and should adhere to safety requirements (for obvious reasons). The proposed project aims to demonstrate how existing safe model-based RL methods can be used to solve the foregoing challenges.
Keywords
safe mode-base RL, online learning, legged robotics
Labels
Master Thesis
Description
Contact Details
More information
Published since: 2024-11-26
Organization ETH Competence Center - ETH AI Center
Hosts Li Chenhao , Li Chenhao , Li Chenhao , Li Chenhao
Topics Engineering and Technology
Autonomous Curriculum Learning for Increasingly Challenging Tasks
While the history of machine learning so far largely encompasses a series of problems posed by researchers and algorithms that learn their solutions, an important question is whether the problems themselves can be generated by the algorithm at the same time as they are being solved. Such a process would in effect build its own diverse and expanding curricula, and the solutions to problems at various stages would become stepping stones towards solving even more challenging problems later in the process. Consider the realm of legged locomotion: Training a robot via reinforcement learning to track a velocity command illustrates this concept. Initially, tracking a low velocity is simpler due to algorithm initialization and environmental setup. By manually crafting a curriculum, we can start with low-velocity targets and incrementally increase them as the robot demonstrates competence. This method works well when the difficulty correlates clearly with the target, as with higher velocities or more challenging terrains. However, challenges arise when the relationship between task difficulty and control parameters is unclear. For instance, if a parameter dictates various human dance styles for the robot to mimic, it's not obvious whether jazz is easier than hip-hop. In such scenarios, the difficulty distribution does not align with the control parameter. How, then, can we devise an effective curriculum? In the conventional RSL training setting for locomotion over challenging terrains, there is also a handcrafted learning schedule dictating increasingly hard terrain levels but unified with multiple different types. With a smart autonomous curriculum learning algorithm, are we able to overcome separate terrain types asynchronously and thus achieve overall better performance or higher data efficiency?
Keywords
curriculum learning, open-ended learning, self-evolution, progressive task solving
Labels
Master Thesis
Description
Contact Details
More information
Published since: 2024-11-26
Organization Robotic Systems Lab
Hosts Li Chenhao , Li Chenhao , Li Chenhao , Bagatella Marco , Li Chenhao
Topics Engineering and Technology
Humanoid Locomotion Learning with Human Motion Priors
Humanoid robots, designed to replicate human structure and behavior, have made significant strides in kinematics, dynamics, and control systems. Research aims to develop robots capable of performing tasks in human-centric settings, from simple object manipulation to navigating complex terrains. Reinforcement learning (RL) has proven to be a powerful method for enabling robots to learn from their environment, enhancing their performance over time without explicit programming for every possible scenario. In the realm of humanoid robotics, RL is used to optimize control policies, adapt to new tasks, and improve the efficiency and safety of human-robot interactions. However, one of the primary challenges is the high dimensionality of the action space, where handcrafted reward functions fall short of generating natural, lifelike motions. Incorporating motion priors into the learning process of humanoid robots addresses these challenges effectively. Motion priors can significantly reduce the exploration space in RL, leading to faster convergence and reduced training time. They ensure that learned policies prioritize stability and safety, reducing the risk of unpredictable or hazardous actions. Additionally, motion priors guide the learning process towards more natural, human-like movements, improving the robot's ability to perform tasks intuitively and seamlessly in human environments. Therefore, motion priors are crucial for efficient, stable, and realistic humanoid locomotion learning, enabling robots to better navigate and interact with the world around them.
Keywords
motion priors, humanoid, reinforcement learning, representation learning
Labels
Master Thesis
Description
Contact Details
More information
Published since: 2024-11-26
Organization ETH Competence Center - ETH AI Center
Hosts Li Chenhao , Li Chenhao , Li Chenhao , Li Chenhao
Topics Information, Computing and Communication Sciences