Student Projects and Jobs

Show initiative!

If no suitable topic or position could be found in this overview, but there is still interest in participating in our chair, please leave a message at the following email address:

work-and-thesis@rsi.ei.tum.de

Setup of a multi-modal avatar station for human-robot interaction in virtual reality

Type: Research Internship

Movies such as Avatar and Ready Player One hypothesized about technology that would connect a human user to a remote body and absolutely immerses them into the experience of the "remote self". Today's real-word technology does not quite achieve this dream but we are currently on our way towards it.

The AI factory is a new paradigm of highly connected, intelligent and modular manufacturing and assembly procedures embedded in completely automated and interactive infrastructure. One of the significant differences to traditional factory setups is that part of the AI factory exists in the virtual world. Human user may work on-site but also from home by use of an avatar station that allows them to connect to robots in the factory's real-world workcells and work alongside other humans and robots from remote. State-of-the-art technologies such as bilateral telepresence, smart wearables and exoskeletons allows to immerse the user by transferring audiovisual and haptic perception in both directions using modern high-speed communication methods such as 5G.

Perfecting the avatar technology would mean a democratization of work, faster reaction times and lower error rates due to highly increased flexibility, availability and expert distribution.

In order to work towards this vision, we intend to setup the first real avatar station in multiple stages. The first stage is the topic of this work and consists more concretely of:

  • Containerizing the available technology
    • Writing standalone programs for different devices
    • Ensure communication between them
    • Devices of stage I include VR glasses, VR gloves, robot, robot hand
  • Setup of dedicated hardware framework
    • Connecting everything
    • Making it "platform-ready" i.e. other technology can be employed on top

Requirements:

  • General affinity towards modern technology
  • Strong background in C++ (Python is a plus)
  • Can work with both Linux and Windows

Nice-to-have:

  • Familiarity with robotics
  • Familiarity with VR technology

Contact:

Lars Johannsmeier, lars.johannsmeier@tum.de

Robot Learning

Requirement Analysis and Evaluation of Reinforcement Learning Methods

With recent advances in machine learning and especifically in reinforcement learning (RL), many new problems have arised.  Specifically, due to the rapid developement of new RL algorithms, not enough efforts have been made into ensuring reproducable results and better benchmarking. This topic is the focus of this internship. We will be conducting different RL experiments to validate different sub-component of the overall pipeline and the effect of small system design details on the overall performance. 

Type: Research/Engineering Internship, IDP, Guided Research

Prerequisites:

  • Studies in electrical, mechanical engineering, computer science or similar study program
  • Solid understanding of basic RL and ML theory
  • Experience/Proficiency with Python (specifically with PyTorch and/or Tensorflow)

Contact: elie.aljalbout@tum.de

Model Learning for physical systems

Model learning is a challenging problem in robotics. When successful it enables robust robot control and long-term planning. While classical methods for system identification have proven to work well in practice, they still lack the autonomy and online features.  One interesting approach for this problem is to use modern machine learning theory e.g. variational inference for intelligent model learning. in this internship you'll be understanding, implementing and evaluating model learning methods for robotics.

Type: Research/Engineering Internship, IDP, Guided Research

Prerequisites:

  • Studies in electrical, mechanical engineering, computer science or similar study program
  • Solid understanding of basic ML theory
  • Experience/Proficiency with Python (specifically with PyTorch and/or Tensorflow)

Contact: carlos.valle@tum.de

Robot Safety: Bachelor/Master Thesis | Research/Engineering-Internship

Designing Collision Test Devices and Collision Testing

Recently the collaboration between robot and human is becoming more and more close. Therefore several constraints for the robot have to be fulfilled. Most importantly human injury has to be prevented. Based on work of Prof. Dr. Haddadin a new portable collision test device has to be build, collision test with human subjects and models for human subjects run and the data merged into safety curves. 
 

Tasks:

- CAD modelling and construction of a collision test device for human limbs 
- Validation of the device and propsed models
- Testing and data collection

Type:

Internship, Bachelor-/Masterthesis
(The tasks can be spllit into diffrent workpackages)

Prerequisites:

- experience in CAD-modelling
- C++ knowledge
- basic biomechanical knowledge

Contact:

robin-jeanne.kirschner@tum.de

Entwicklung von Testaufbauten für Kollisionsexperimente in der Robotik

Im industriellen und häuslichen Umfeld werden Roboter zukünftig eng mit dem Menschen zusammenarbeiten. Während der physischen Interaktion kann grundsätzlich nicht vermieden werden, dass unerwünschter Kontakt auftritt. Um die Sicherheit des Menschen zu gewährleisten, muss untersucht werden, was passieren kann, wenn Mensch und Roboter miteinander kollidieren. An der
 Munich School of Robotics der TU München soll ein Labor aufgebaut werden, in dem systematische Kollisionsexperimente durchgeführt werden, um die vom Roboter ausgehende Gefahr beurteilen zu können. Folgende Themen können im Rahmen der studentischen Arbeit behandelt werden:

  • Konzeption und Planung verschiedener Teststände

  • CAD-Design der Aufbauten

  • Auslegung der Elektronik, Sensorik und Aktuierung

  • Integration der Komponenten

  • Entwicklung der Steuerung

  • Durchführung von Experimenten und Auswertung der Daten

 

Voraussetzungen

  • Studium der Elektrotechnik, Maschinenbau, Mechatronik, Informatik o.ä.

  • Eigenständige Arbeitsweise und Kreativität

 

In der Arbeit sind verschiedene Schwerpunkte möglich, eine konkrete Aufgabenstellung kann in einem persönlichen Gespräch erörtert werden.

Praktikanten- und Hilfswissenschaftler-Stellen sind besonders gesucht!


 Kontakt

Robin Kirschner, M.Sc.

Munich School of Robotics and Machine Intelligence
 Technische Universität München
 Heßstraße 134
 80797 München

robin-jeanne.kirschner@tum.de
 +49 (89) 289 – 29438

Collision Analysis and Safe Control in Human-Robot Interaction

Currently, increasing effort is taken in the robotics community to understand injury mechanisms during physical human-robot interaction (pHRI). This is motivated by the fact that human and robot will work intensively and closely together, and therefore, one has to be aware of the potential threats in case such a close cooperation takes place and take appropriate countermeasures to ensure human safety via planning and/or control. In the context of safety in pHRI, possible topics that can be addressed in the thesis/internship are:

  • Design and analysis of collision experiments and/or testing devices
  • Development and verification of collision simulations
  • Survey of biomechanics and forensics literature
  • Development of robot motion planning and/or control schemes for ensuring human safety

Prerequisites:

  • Studies in Mechanics, Mechatronics, Electronics, Computer Science 
  • Knowledge in robotics & control (for topics on planning & control)
  • Good C++ programming skills
  • Matlab/Simulink
  • Working knowledge in ROS
  • Ability to work well structured and organized
  • Creativity

Contact:
Mazin Hamad, M.Sc.
mazin.hamad@tum.de
Chair of Robotics Science and Systems Intelligence

Analysis and safe control of mobile manipulators

Mobile robots have become increasingly popular in industrial and service scenarios, as they allow for dexterous motions and manipulation in large areas. As robots are nowadays intended to work closely together with human operators, one needs to be aware of potential threats and take appropriate countermeasures to ensure human safety via planning and control. In this work, the safety characteristics of mobile manipulators shall be investigated and control schemes shall be developed which maximize performance while meeting safety constraints at the same time. 

Possible work steps:

  • Literature review
  • Modelling and comparison of different types of mobile platforms
  • Derivation of global safety characteristics
  • Development of safe control schemes

Prerequisites:

  • Studies in Mechanics, Mechatronics, Electronics, or Computer Science 
  • Knowledge in robotics & control
  • Knowledge in optimization
  • Matlab/Simulink
  • Ability to work well structured and organized
  • Creativity

Contact:
Nico Mansfeld, M.Sc.
nico.mansfeld@tum.de

Alexander Kurdas, M.Sc.
alexander.kurdas@tum.de

Chair of Robotics Science and Systems Intelligence

Artificial Hand Design

Advance and Improvement of an Artificial Hand Testbed

→ Engineering / Research Internship
→ For more information contact us!

Prerequisites:

  • CAD (e.g. Solidworks)
  • Design and Production methods construction parts
  • Hardware implementation

Contact:
Johannes Ringwald
johannes.ringwald@tum.de
+49 (89) 289 - 29414

Task Planning

AI-Enabled Lab-Automation

AI-supported robotic Lab Assistant: Reliable automated liquid handling for the daily laboratory routine of the future

Type: Research/Engineering-Internship

→ For more information contact us!

The automation of laboratory processes in chemistry, biotechnology, pharmaceuticals, food technology and medicine is already a reality today. However, many solutions offered on the market today are either too expensive and/or only specially developed and optimized for certain processes in the laboratory. In order that laboratory automation can be used as a tool for everyone in the daily laboratory routine, an automation process for a AI-supported robotic lab assistant is to be developed and implemented in the laboratory in this application research project. As a process to be automated, reliable automated liquid handling is to be tackled for the time being. To achieve this, one of the latest collaborative robots is available as well as 3D printing options for flexible gripper finger development.

 

The following procedure is planned:

  • Extensive literature research Robotics in laboratory automation
  • Analysis and development of robotic finger systems for the use of laboratory tools
  • Analysis, automation and evaluation of process sequences with regard to liquid handling in the laboratory
  • And much more ...

Prerequisites:

  • From the fields of Electrical Engineering and Information Technology, Mechanical Engineering or Mechatronics
  • Basic knowledge in robotics, control engineering and system theory
  • Good programming skills in C/C++, Python, Matlab
  • Good CAD design skills (SolidWorks etc.)
  • Experience with 3D printing
  • Creative but structured and independent work

Contact:
Dennis Knobbe
dennis.knobbe@tum.de
+49 (89) 289 -29412

Optimal Sensor Configuration

Optimal Sensor Configuration

Sensor fusion and filtering is an important topic in signals and systems as well as robotics. More specifically, combining sensors of different types enables one for further robot analysis and devise advanced control strategies. However, in order to get optimal technical and computation costs, the number of sensors has to be carefully chosen. Moreover, for some sensors such as accelerometers whose output signal depends on their location, the sensor placement needs to be optimally chosen, as well. Within the framework of these criteria, we seek to formulate an algorithm for proper selection of sensors and analyze the method optimality, in this thesis.

Who Are You?

  • An enthusiastic mechanical/electrical/robotic master student
  • Interested in signals and systems analysis and optimization problems
  • Familiar with sensor technologies
  • Familiar with Matlab, Simulink and python

Contact

ali.baradaran@tum.de

Human modeling

Teleoperation

Development of a telepresence reference platform

Type: Research internship

Tasks

The goal is to develop and setup a telepresence reference platform consisting of state-of-the-art robots to objectively compare teleoperation control schemes and communication methodologies. Concrete work packages are:

  • Planning and setup of hardware consisting of table, robots and additional items.
  • Development of objective evaluation procedures including quality metrics.
  • Initial development of user studies.

Requirements

  • Experienced with C++ (this is very important)
  • Familiar with python
  • Basic knowledge in robotics
  • Familiar with (at least interested in) telepresence

Motivation

Teleoperation will become more important in the next years due to rapid advances in modern communication technologies such as 5G. The next stage of the internet, the Tactile Internet, will allow user to interact with remote location via robotic avatars. High-speed communciation relays not only audiovisual information but also tactile feedback to the user which allows for completely new applications.

Although there has been much research into robotic telepresence and communication modelling over the last decades, we are often still lacking objective criterions to compare the various approaches and combinations of them. The aim of this research internship is to develop a platform consisting of state-of-the-art robots that can serve as a basis for experimental comparisons. Building on this, objective quality metrics based on both robotic and human interaction with the system will be investigated.

Please also have a look at this paper for further reference.

Contact

Lars Johannsmeier

lars.johannsmeier@tum.de