Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
Next revision		Previous revision
Next revisionBoth sides next revision
jobs [2014/02/18 17:01] – tenorthjobs [2016/02/24 13:16] – gkazhoya
Line 3: Line 3:
 If you are looking for a bachelor/master thesis or a job as a student research assistant, you may find some interesting opportunities on this page. If you are looking for a bachelor/master thesis or a job as a student research assistant, you may find some interesting opportunities on this page.
  
-====Theses====+== Lisp / CRAM support assistant (HiWi) ==
  
-== Web-based knowledge visualizations (BA/HiWi) ==+Technical support for the group for Lisp and the CRAM framework. \\ 
 +5 hours per week for up to 1 year (paid).
  
-{{ :research:knowledge-visualization.png?200}}+Requirements: 
 +  * Good programming skills in Common Lisp 
 +  * Basic ROS knowledge 
 + 
 +The student will be introduced to the CRAM framework at the beginning of the job, which is a robot programming framework written in Lisp. The student will then be responsible for assisting not familiar with the framework people, explaining them the parts they don't understand and pointing them to the relevant documentation sources. 
 + 
 +Contact: [[team:gayane_kazhoyan|Gayane Kazhoyan]] 
 + 
 + 
 +== Integrating PR2 in the Unreal Game Engine Framework (BA)== 
 + {{ :research:unreal_ros_pr2.png?200|}} 
  
-The knowledge base visualization allows to display semantic environment maps, +Integrating the [[https://www.willowgarage.com/pages/pr2/overview|PR2]] robot with [[http://www.ros.org/|ROS]] support in the [[https://www.unrealengine.com|Unreal Engine 4]] Framework.
-objects, human poses, trajectories and similar information stored in the +
-robot's knowledge base. Its [[http://www.ros.org/wiki/mod_vis|current version]] +
-is written in Java and has grown old over the years. Recently, there has been +
-much progress in creating [[http://www.ros.org/wiki/interactive_markers|interactive +
-3D visualizations]] for ROS that can also be accessed with a +
-[[http://www.ros.org/wiki/wviz|web browser]]. This project is about creating +
-a web-based version of the visualization using these techniques.+
  
 Requirements: Requirements:
-  * Good programming skills +  * Good programming skills in C/C++ 
-  * Experience with JavaScript development +  * Basic physics/rendering engine knowledge 
-  * Knowledge of Web technologies (HTML, XML, OWL)+  * Basic ROS knowledge 
 +  * UE4 basic tutorials
  
-Contact: [[team:moritz_tenorth|Moritz Tenorth]]+Contact: [[team:andrei_haidu|Andrei Haidu]]
  
 +== Kitchen Activity Games in a Realistic Robotic Simulator (BA/MA)==
 + {{ :research:gz_env1.png?200|}} 
  
-== Statistical relational models of human activities (MA) ==+Developing new activities and improving the current simulation framework done under the [[http://gazebosim.org/|Gazebo]] robotic simulator. Creating a custom GUI for the game, in order to launch new scenarios, save logs etc.
  
-As part of the [[http://cogwatch.eu|CogWatch EU project]], we are working on +Requirements: 
-the analysis of observations of Apraxia patients. After suffering brain damage, +  * Good programming skills in C/C++ 
-these patients lack important motor coordination skills that result in various +  * Basic physics/rendering engine knowledge 
-kinds of errors in performing everyday activities. Based on data from the +  * Gazebo simulator basic tutorials 
-extensive patient trials, the student is to learn statistical relational + 
-models that characterize which errors are made in which circumstances.+Contact: [[team:andrei_haidu|Andrei Haidu]] 
 + 
 +== Integrating Eye Tracking in the Kitchen Activity Games (BA/MA)== 
 + {{ :research:eye_tracker.png?200|}}  
 + 
 +Integrating the eye tracker in the [[http://gazebosim.org/|Gazebo]] based Kitchen Activity Games framework and logging the gaze of the user during the gameplay. From the information typical activities should be inferred.
  
 Requirements: Requirements:
-  * Good knowledge of machine learning (Bayesian networks, statistical relational models) +  * Good programming skills in C/C++ 
-  * Programming skills (Java)+  * Gazebo simulator basic tutorials
  
-Contact: [[team:moritz_tenorth|Moritz Tenorth]]+Contact: [[team:andrei_haidu|Andrei Haidu]]
  
 +== Hand Skeleton Tracking Using Two Leap Motion Devices (BA/MA)==
 + {{ :research:leap_motion.jpg?200|}} 
  
-== Tools for knowledge acquisition from the Web (BA/MA/HiWi) ==+Improving the skeletal tracking offered by the [[https://developer.leapmotion.com/|Leap Motion SDK]], by using two devices (one tracking vertically the other horizontally) and switching between them to the one that has the best current view of the hand.
  
-There are several options for doing a project related to the acquisition of +The tracked hand can then be used as input for the Kitchen Activity Games framework.
-knowledge from Web sources like online shops, repositories of object models, +
-recipe databases, etc.+
  
 Requirements: Requirements:
-  * Programing skills (Java) +  * Good programming skills in C/C++
-  * Experience with Web languages and datamining techniques is helpful +
-  * Depending on the focus of the project, experience with database technology, natural-language processing or computer vision may be helpful+
  
-Contact: [[team:moritz_tenorth|Moritz Tenorth]]+Contact: [[team:andrei_haidu|Andrei Haidu]]
  
 +== Fluid Simulation in Gazebo (BA/MA)==
 + {{ :research:fluid.png?200|}} 
  
-== GPU-based Parallelization of Numerical Optimization Techniques (BA/MA/HiWi)==+[[http://gazebosim.org/|Gazebo]] currently only supports rigid body physics engines (ODE, Bullet etc.), however in some cases fluids are preferred in order to simulate as realistically as possible the given environment.
  
-In the field of Machine Learning, numerical optimization techniques play a focal role. However, as models grow larger, traditional implementations on single-core CPUs suffer from sequential execution causing a severe slow-down. In this thesis, state-of-the-art GPU frameworks (e.g. CUDA) are to be investigated in order implement numerical optimizers that substantially profit from parallel execution.+Currently there is an [[http://gazebosim.org/tutorials?tut=fluids&cat=physics|experimental version]] of fluids  in Gazebo, using the [[http://onezero.ca/fluidix/|Fluidix]] library to run the fluids computation on the GPU.
  
-Requirements: +The computational method for the fluid simulation is SPH (Smoothed-particle Dynamics), however newer and better methods based on SPH are currently present 
-  * Skills in numerical optimization algorithms +and should be implemented (PCISPH/IISPH).
-  * Good programming skills in Python and C/C+++
  
-Contact: [[team:daniel_nyga|Daniel Nyga]]+The interaction between the fluid and the rigid objects is a naive one, the forces and torques are applied only from the particle collisions (not taking into account pressure and other forces).
  
-== Online Learning of Markov Logic Networks for Natural-Language Understanding (MA)==+Another topic would be the visualization of the fluid, currently is done by rendering every particle. For the rendering engine [[http://www.ogre3d.org/|OGRE]] is used.
  
-Markov Logic Networks (MLNs) combine the expressive power of first-order logic and probabilistic graphical models. In the past, they have been successfully applied to the problem of semantically interpreting and completing natural-language instructions from the web. State-of-the-art learning techniques mostly operate in batch mode, i.e. all training instances need to be known in the beginning of the learning process. In context of this thesis, online learning methods for MLNs are to be investigated, which allow incremental learning, when new examples come in one-by-one.+Here is a [[https://vimeo.com/104629835|video]] example of the current state of the fluid in Gazebo. 
  
 Requirements: Requirements:
-  * Experience in Machine Learning. +  * Good programming skills in C/C++ 
-  * Experience with statistical relational learning (e.g. MLNs) is helpful. +  * Interest in Fluid simulation 
-  * Good programming skills in Python.+  * Basic physics/rendering engine knowledge 
 +  * Gazebo simulator and Fluidix basic tutorials
  
-Contact: [[team:daniel_nyga|Daniel Nyga]]+Contact: [[team:andrei_haidu|Andrei Haidu]]
  
  
-<html> <br> </br/> </br/> </html> +== Automated sensor calibration toolkit (BA/MA)==
-====Open Positions==== +
-==HiWi-Position: Knowledge Representation & Language Understanding for Intelligent Robots==+
  
-In the context of the European research project RoboHow.Cog [1,2] we +Computer vision is an important part of autonomous robots. For robots the image sensors are the main source of information of the surrounding world. Each camera is different, even if they are from the same production line. For computer vision, especially for robots manipulating their environment, it is important that the parameters for the cameras in use are well known. The calibration of a camera is a time consuming task, and the result depends highly on the chosen setup and the accuracy of the operator.
-are investigating methods for combining multimodal sources of knowledge (e.g. video, natural-language recipes or computer games), in order to enable mobile robots to autonomously acquire new high level skills like cooking meals or straightening up rooms.  +
  
-The Institute for Artificial Intelligence is hiring a student researcher for the +The topic for this thesis is to develop an automated system for calibrating cameras, especially RGB-D cameras like the Kinect v2.
-development and the integration of probabilistic methods in AI, which enable intelligent robots to understand, interpret and execute natural-language instructions from recipes from the World Wide Web.+
  
-This HiWi-Position can serve as a starting point for future Bachelor's, Master's or Diploma Theses.+ {{ :kinect2_calibration_setup_small.jpg?200|}} 
 +The system should: 
 +  * be independent of the camera type 
 +  * estimate intrinsic and extrinsic parameters 
 +  * calibrate depth images (case of RGB-D) 
 +  * integrate capabilities from Halcon [1] 
 +  * operate autonomously
  
-Tasks: +Requirements:  
-  * Implementation of an interface to the Robot Operating System (ROS). +  * Good programming skills in Python and C/C++ 
-  * Linkage of the knowledge base to the executive of the robot. +  * ROS, OpenCV
-  * Support for the scientific staff in extending and integrating components onto the robot platform PR2.+
  
-Requirements: +[1] http://www.halcon.de/
-  * Studies in Computer Science (Bachelor's, Master's or Diploma) +
-  * Basic skills in Artificial Intelligence +
-  * Optional: basic skills in Probability Theory +
-  * Optional: basic skills in Machine Learning +
-  * Good programming skills in Python and Java+
  
-Hours: 10-20 h/week+Contact: [[team:alexis_maldonado|Alexis Maldonado]] and [[team:thiemo_wiedemeyer|Thiemo Wiedemeyer]]
  
-Contact: [[team:daniel_nyga|Daniel Nyga]]+== On-the-fly 3D CAD model creation (MA)==
  
-[1] www.robohow.eu\\ +Create models during runtime for unknown textured objets based on depth and color information. Track the object and update the model with more detailed information, completing it's 3D model from multiple views improving redetection. Using the robots manipulator pick up the object and complete the model by viewing it from multiple viewpoints.
-[2] http://www.youtube.com/watch?v=0eIryyzlRwA+
  
-== HiWi position: Segmentation and interpretation of 3D object models (HiWi/MA) ==+Requirements:  
 +  * Good programming skills in C/C++ 
 +  * strong background in computer vision  
 +  * ROS, OpenCV, PCL
  
-{{ :research:cup2-segmented.png?170}}+Contact: [[team:thiemo_wiedemeyer|Thiemo Wiedemeyer]]
  
-Competent object interaction requires knowledge about the structure and +== Simulation of a robots belief state to support perception(MA) ==
-composition of objects. In an ongoing research project, we are investigating +
-how part-based object models can automatically be extracted from CAD models +
-found on the Web, e.g. on the [[http://sketchup.google.com/3dwarehouse|3D warehouse]].+
  
-We are looking for a student research assistant to push this topic forward. +Create a simulation environment that represents the robots current belief state and can be updated frequently. Use off-screen rendering to investigate the affordances these objects possess, in order to support segmentation, detection and tracking of these in the real world. 
-In close collaboration with the researchers of the Institute for Artificial +
-Intelligence, the student will work on extending the software for mesh +
-segmentation, implement novel algorithms, improve the integration of the +
-methods into the robot's knowledge base, and update the modules for visualizing +
-the segmentation results. Depending on the results, cooperation on joint  +
-publications may be possible.+
  
-This HiWi position can serve as a starting point for future Bachelor's, +Requirements:  
-Master's or Diploma Theses. Alternatively, we could extract a research +  * Good programming skills in C/C++ 
-project that could be worked on as a Master's or Diploma Thesis.+  * strong background in computer vision  
 +  * Gazebo, OpenCV, PCL
  
-Requirements: +Contact: [[team:ferenc_balint-benczedi|Ferenc Balint-Benczedi]] 
-  * Good Java programming skills + 
-  * Basic knowledge of 3D geometry calculations +== Multi-expert segmentation of cluttered and occluded scenes == 
-  * Experience with working with 3D models is helpful+ 
 +Objects in a human environment are usually found in challenging scenes. They can be stacked upon eachother, touching or occluding, can be found in drawers, cupboards, refrigerators and so on. A personal robot assistant in order to execute a task, needs to detect these objects and recognize them. In this thesis a multi-modal approach to interpreting cluttered scenes is going to be investigated, combining the results of multiple segmentation algorithms in order to come up with more reliable object hypotheses. 
 + 
 +Requirements:  
 +  * Good programming skills in C/C++ 
 +  * strong background in 3D vision  
 +  * basic knowledge of ROS, OpenCV, PCL 
 + 
 +Contact: [[team:ferenc_balint-benczedi|Ferenc Balint-Benczedi]]
  
-Contact: [[team:moritz_tenorth|Moritz Tenorth]] 
  
- Impressum - Datenschutz



Prof. Dr. hc. Michael Beetz PhD
Head of Institute

Contact via
Andrea Cowley
assistant to Prof. Beetz
ai-office@cs.uni-bremen.de

Discover our VRB for innovative and interactive research


Memberships and associations:


Social Media: