![](/files/organization-images/9043abb3-87e6-4268-863b-36fc497d5444/0f1e63c5-e610-44c2-8ca9-d2931f987950.png_200_200.png) pd|z Product Development Group ZurichOpen OpportunitiesOur current action primitive framework characterizes the underlying primitives by a set of Impedance parameters (safety bubble, robot control behaviour). However, those are fixed during runtime and cannot be changed. We would like to explore the possibilities of MPC-based online adaptation of the robot’s virtual stiffness, to maximize performance, while retaining safety and interaction modality constraints. For this we need to consider a model of the human movement and formulate adequate constraints for our optimization problem. Furthermore, stability issues during online parameter changes may arise, that need to be taken care of. At the end, we would like to show the validity of this approach with a user study, comparing it especially to our old (fixed-parameter approach). - Engineering and Technology
- Master Thesis
| The goal of this thesis is to implement a method that allows for efficient handovers between the collaborative robots and humans. To this end you should leverage state-of-the-art computer vision and AI methods to generate a model of the human’s hand. Then we need to create an algorithm that determines the optimal approach to the hand. Finally, we need to account for interaction object geometries. This is the pipeline in a static case: However, when the user is moving, we also do not know, where the handover will take place. Thus, this thesis could take one of two main directions: - Engineering and Technology
- Master Thesis
| The goal of this thesis is to achieve a proactive human-robot collaboration (HRC) using real-time sensor measurements from the user. Currently, we implemented a HRC system that enables different collaborative tasks (object handovers, hand following, etc.). However, at the moment this system is purely reactive and requires gestures or speech commands to interact with the robot. Thus, we would like to implement a human intent prediction, based on sensor measurements, such as skeleton tracking, gaze tracking, IMU measurements or even heartrate and blood pressure (through a smartwatch). This could be based on AI or other probabilistic models, where the system proactively prepares the next action or even pre-empts it entirely. This would require a series of case studies to gather data about human behavior during interaction with our system, with subsequent data analysis and training or fitting of the models. The whole Pipeline should be demonstrated on a select use-case.
- Engineering and Technology
- Master Thesis
| In collaborative robotics two main control schemes stand at the center of research, as they allow to render comliant robot behaviour when subject to external forces: Impedance and admittance control. They can be seen as complementary methods (the robot being modelled as a mechanical impedance or admittance respectively while the environment is the inverse). However, they have very distince interaction and stability properties. While admittance control generally allows for better performance when the environment has a low stiffness (e.g. free-floating motion), Impedance control is better suited to environments with high stiffness, as the former exhibits stabilty issues in this case, while the latter lacks precision when an imperfect model of the robot dynamics is used. Some Authors have thus already proposed hybrid schemes, whereby a continuous change from impedance to admittance control can be commanded at runtime. - Engineering and Technology
- Bachelor Thesis, Master Thesis, Semester Project
| Current 3D perception pipelines severely lack in accuracy and performance. Inherent noise in Point Cloud measurements, as well as occlusions starts one off with sub-par data. Additionally, there is very little annotated data available for direct Point cloud segmentation. Thus, workarounds have been tested, like depth projection of 2D segmentation masks [us, SAMPRO3D…]. However, they tend to be slow, because of the need of various views to reconstruct the scene, with additional cameras. Furthermore, they require previous semantic segmentation. Direct Point cloud segmentation has the potential to be much faster, since multiple view angles can easily be concatenated. However, they lack the right sizes and quality of datasets to build foundational models. Your task would thus be to finetune or create a Neural Network for Point cloud segmentation, as well as a dataset for supervised learning. For this, you can use our preexisting vision pipelines or data available online. To create annotations, we propose to automatically generate ground-truth labels with SAM-Pro 3D to keep manual labelling minimal. - Engineering and Technology
- Master Thesis
| One of the most challenging tasks in HRC with Impedance control is the avoidance of singularities, as they promote oscillations, erratic behaviour and very fast joint motion. Normally, they are circumvented through adequate path planning or pre-programming of robot trajectories. However, during interactive tasks, the robot might be subject to dynamic external forces (through direct interaction or safety frameworks for collision avoidance) which deviate the end-effector from the planned path. Your goal is to develop a singularity avoidance algorithm, which stops the robot from moving into these configurations, while (in a reasonable manner) not blocking the entire task execution. This may involve path re-planning, null space projection, virtual torques or any other method you might find suited to the task. - Engineering and Technology
- Bachelor Thesis, Master Thesis, Semester Project
| The gravity compensation model provided by the manufacturer of our robot is not precise enough: during free-floating the end-effector might fall a few centimetres, before coming to a halt, which we suspect might only be through friction. Your task is to conceive a system-Identification procedure that allow us to calibrate the robot’s Coriolis, friction and gravity torques at will. You should validate your approach by comparing your values with the baseline from the robot manufacturer. - Engineering and Technology
- Bachelor Thesis, Master Thesis, Semester Project
| Drying (e.g. Pasta drying) is the most energy intensive process step, sometimes taking up more than 50% of the total energy consumption of a plant. Superheated steam drying could present an energy efficient alternative to classical hot-air drying systems used today. This new technology could have a massive impact on the carbon-footprint and sustainability of food-drying; making it a highly future-oriented and potentially impactful innovation. - Interdisciplinary Engineering, Manufacturing Engineering, Mechanical and Industrial Engineering
- ETH Zurich (ETHZ), Master Thesis, Semester Project
| This project focuses on enhancing the accuracy and efficiency of CT-guided medical interventions through the development of an Augmented Reality (AR) application for the Apple Vision Pro. By integrating advanced AR technologies with the Cube Navigation System (CNS), the project aims to improve needle placement precision, reducing procedural errors and intervention time. The development process will involve creating a sophisticated 3D environment and leveraging cutting-edge tracking and deep learning algorithms. - Clinical Engineering, Image Processing
- Master Thesis
| The aim of this project is to evaluate surgeons' skills using movement data from surgical tools through advanced artificial intelligence methods for sequential data analysis. - Biomedical Engineering, Image Processing, Signal Processing
- Master Thesis
|
|