Student Projects
Review the projects below and send an email to the project contact.
Teaching a robot to master a board game using Reinforcement Learning
The master thesis aims to create a physics-based simulation with Brax or MuJoCo for a robot to learn board game strategies via reinforcement learning. Utilizing a pre-built 2x 2D gantry robot integrated with ROS2, the project seeks to train the robot to learn the game from scratch and improve its play against both human and robotic opponents. The primary goal is rapid learning and superior performance, with a focus on understanding the factors of victory in games like this between robots trained on the same algorithm. All this will be performed in simulation and on a real robot.
Labels
Semester Project , Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-10-28 , Earliest start: 2024-08-15
Organization Research D'Andrea
Hosts Ramachandran Aswin
Topics Engineering and Technology
One Step Diffusion via First-Order Optimality Conditions in the Wasserstein Space
The stochastic diffusion equations ruling the dynamics of particles at the micro- and nano- scale are captured by energy-minimizing dynamics when observed macroscopically, i.e., at a population level. This framework encompasses, for instance, single cells perturbation responses to chemical, genetic or mechanical stimuli, gene expression and cell differentiation. In our recent work [1], we deployed a theory we developed for studying optimization problems in the Wasserstein space [2] to speed up the learning of the different terms of these diffusion processes from observational data. In this work, we will instead focus on how to deploy these tools to distill diffusion models efficiently for one-step image generation and robotics control. The thesis will validate the developed method on real-world robotics setup available and ready to use. References: [1] Learning diffusion at lightspeed. A. Terpin, N. Lanzetti, M. Gadea, F. Dörfler. NeurIPS 2024 (oral) [2] Variational Analysis in the Wasserstein Space. N. Lanzetti*, A. Terpin*, F. Dörfler. Pre-print.
Labels
Semester Project , Master Thesis
More information
Published since: 2024-10-20 , Earliest start: 2024-10-21
Organization Research D'Andrea
Hosts Terpin Antonio
Topics Information, Computing and Communication Sciences
Tech meets Art: Design and development of a "magic box"
Description: When carefully combined, mathematics, control theory, and engineering are capable of creating technology undistinguishable from magic. In this project, we explore a design for a small box (more details will be provided to matching applicants) that exploits principles of control theory, electro-magnetism, and vision to provide a magic-like effect on the observer. You will take responsibility (supported by the researchers and technical staff at IDSC) of one area of development of this "magic box": - Actuation: combining control theory and electromagnetism, you will design and implement a very compact and noiseless linear system with minimal yet sufficient range of motion. - Vision: you will implement very efficient and effective vision algorithms to empower the magic box with decision capabilities, using time-of-flights sensors. - Design: You will work on the aesthetics and mechanical design of the overall box, ensuring compatibility with the rest of the components.
Labels
Semester Project , Master Thesis
Contact Details
More information
Published since: 2024-10-20 , Earliest start: 2024-10-21
Organization Research D'Andrea
Hosts Terpin Antonio
Topics Engineering and Technology
Design and optimization of a table-top water tunnel
In this project, we will optimize a work-in-progress prototype of a small testbed for fluids control problems. Your task will be to model the setup in a computational fluid dynamics simulator, and optimize the various parts to improve the flow quality. The proposed modifications will be manufactured with the IDSC technical staff. Differently from most water tunnels, our design is table top and thus presents strong spatial constraints. The project will create detail technical documentation to make the setup easily reproducible by other research groups. If you are excited about the challenge, email your application with your portfolio and transcript. We are looking for a student with experience in CAD design, computational fluid dynamics, Ansys or similar modeling and simulation software, and prototyping.
Keywords
fluiddynamics, computational fluid dynamics, water tunnel, design, optimization
Labels
Semester Project , Bachelor Thesis , Master Thesis
PLEASE LOG IN TO SEE DESCRIPTION
More information
Published since: 2024-09-17 , Earliest start: 2024-10-01
Organization Research D'Andrea
Hosts Terpin Antonio
Topics Engineering and Technology
The Way of Water: Development of a fleet of water-based drones for live performance
This project focuses on developing autonomous robots for synchronized performances on water. Equipped with kinetic water fountains, RGB lighting, and ultrasonic mist generators, the robots are designed to execute planned choreographies. The system utilizes robotics control, wireless communication, and positioning technologies to coordinate movements, and payload activation, facilitating complex pattern generation and synchronization. The objective is to advance the application of distributed robotic systems in creating structured and cohesive visual displays on water.
Keywords
Water based rovers, Electronics development, Distributed Robotics, Control Systems, Game design
Labels
Semester Project , Bachelor Thesis , Master Thesis
Description
Contact Details
More information
Published since: 2024-08-24
Organization Research D'Andrea
Hosts Ramachandran Aswin
Topics Arts , Information, Computing and Communication Sciences , Engineering and Technology
CyberRunner: Pushing the Boundary of Autonomy
CyberRunner is an AI robot whose task is to learn how to play the popular and widely accessible labyrinth marble game. The labyrinth is a game of physical skill whose goal is to steer a marble from a given start point to the end point. In doing so, the player must prevent the ball from falling into any of the holes that are present on the labyrinth board. The movement of the ball can be indirectly controlled by two knobs which change the orientation of the board. While it is a relatively straightforward game, it requires fine motor skills and spatial reasoning abilities, and, from experience, humans require a great amount of practice to become proficient at the game. Using recent advances in model-based reinforcement learning techniques, CyberRunner is able to outperform the previously fastest recorded time, achieved by an extremely skilled human player, by over 6%. Moreover, it does so with only 6 hours of practice. We envision expanding the capabilities of CyberRunner through further research. Students will contribute to advancing the field and establishing CyberRunner as a real-world robotic benchmark. Suitable projects in different areas are available for talented and motivated students. The project topics span model-based control, reinforcement learning, computer vision, and hardware design.
Keywords
CyberRunner, Reinforcement Learning, Machine Learning, Robotics
Labels
Semester Project , Master Thesis
Contact Details
More information
Published since: 2024-07-24 , Earliest start: 2024-09-01 , Latest end: 2025-09-01
Organization Research D'Andrea
Hosts Bi Thomas
Topics Engineering and Technology