|
|
Dr. Stulp, Freek
Researcher
Postdoctoral researcher
|
Table of ContentsResearch TopicsCurrent Research Projects Former Research Projects Publications |
Welcome to my homepage at the Intelligent Autonomous Systems Group at the Technische Universität München. On this homepage, you will find information on my research projects, what I write about it, and what others write about it. It also lists the teaching I do, as well as miscellaneous other activities.
(Our group has recently switched to a new website layout. Apologies for any broken links or mixed styles)
Research: My research focuses on using action models to generate and optimize sequences of actions for robots. Robots acquire action models by interacting with the world, and generalizing over their experiences with machine learning.
As these machine learning algorithms have proven to be very powerful in the context of face model fitting, this has become a second main thread in my research.
Bio: (My full CV is here) I received the Doctorandus (=MSc.) degree in Cognitive Science and Engineering (now renamed to Artificial Intelligence) at the University of Groningen, where I also studied Biology for two years (as prerequisite for CS&E). During my studies, I held Research Assistant positions at the Institute of Perception, Action and Behaviour at the University of Edinburgh, as well as the Instituto de Sistemas e Robótica at the Instituto Superior Técnico. Afterwards, I conducted my PhD research at the Intelligent Autonomous Systems Group at the Technische Universität München, under the supervision of Michael Beetz. I completed my PhD in October 2007. In 2007, I was a post-doc at the Cognitive Neuroinformatics Group at the University of Bremen. In 2008, I returned to Munich to join the CogMan project ("Cognitive Models of Everyday Manipulation Tasks"), part of the CoTeSys cluster of excellence. In 2009, I was also a Postdoctoral Fellow at the Advanced Telecommunications Research Institute International in Kyoto, Japan
Research Topics
- Motion Primitives
- Human and Robot Manipulation
- Human Motions as Model for Robot Controllers
- Plan Optimization
- Human Motion Capture, Tracking and Analysis
- Human Motions in Manipulation Tasks
- Objective Functions, Learning Objective Functions
Current Research Projects
- CogMan : The CogMan project (1) develops computational and control models of pick-and-place tasks in the context of everyday manipulation activities in human environments, (2) implements the model into a control system for the kitchen scenario, and (3) empirically analyzes the impact of this control model on the flexibility, robustness, adaptability, and naturality of the robot behavior.
- Face Image Analysis : Recent progress in the field of Computer Vision allows intuitive interaction via gesture or facial expressions between humans and technical systems. This project aims at determining facial expressions from camera images in real-time. Model-based image interpretation techniques have proven to be a successful method for extracting such high-level information from single images and image sequences. Models rely on a priori knowledge to reduce the large amount of image data to a small number of model parameters. This representation of the image content facilitates and accelerates the subsequent interpretation task. In order to extract high-level information, model parameters have to be estimated that best describe the face within a given image. However, correctly estimated model parameters forms the basis of various more applications such as gaze detection or gender estimation. Our demonstrator for facial expression estimation has been presented at several events with political audience and on TV.
Former Research Projects
- Agilo : Robotic soccer has become a standard 'real-world' testbed for autonomous multi robot control. In robot soccer (mid-size league) two teams of four autonomous robots --- one goal keeper and three field players --- play soccer against each other. The soccer field is four by nine meters big. The key characteristics of mid-size robot soccer are that the robots are completely autonomous. Consequently, all sensing and all action selection is done onboard of the individual robots. Skillful play requires our robots to recognize objects, such as other robots, field lines, and goals, and even entire game situations. In the AGILO project we investigate how probabilistic visuomotoric autonomous robot controllers that are capable of learning can meet these challenges. The AGILO robot controllers employ game state estimation and situated action selection based on automatically learned control mechanisms.
FAQ
Freekuently Asked Question: The not uncommon dutch name "Freek" is pronounced "Frrrake" (with a Spanish-style "r" as in "burro"). Don't worry, I'm not fussy about this: over the years abroad I have also learned to respond to "Fred", "Vague", "Flake", "Flag", "Freoch", and "Freak" (actually, one of the institutes I worked at always addressed letters to "Dr. Freak", though at the time I was neither). Contrary to the well-known English meaning, the original lesser-known Dutch meaning is "powerful protector". For completeness: "Stulp" means "humble abode".
Selected Publications
-
Action-Related Place-Based Mobile Manipulation, 2009, Proceedings of the International Conference on Intelligent Robots and Systems (IROS). To appear.,(BibTeX)
-
Refining the execution of abstract actions with learned action models, 2008, Journal of Artificial Intelligence Research (JAIR),(BibTeX)
-
Combining Declarative, Procedural and Predictive Knowledge to Generate and Execute Robot Plans Efficiently and Robustly, 2008, Robotics and Autonomous Systems Journal (Special Issue on Semantic Knowledge),(BibTeX)
-
Learning Local Objective Functions for Robust Face Model Fitting, 2008, IEEE Transactions on Pattern Analysis and Machine Intelligence~(PAMI),(BibTeX)
