|Subprojekt||Title||Project Leader, Institute||More information|
|L1||Learning of Practical Knowledge, Episodes, and Context||Prof. Dr.-Ing. Rüdiger Dillmann, Prof. Dr.-Ing. Tanja Schultz||Show more|
Subproject L1 researches the question of how knowledge for a humanoid robot can be learned from humans. External, as well as internal observations of the human are investigated. On the one hand we acquire new task knowledge by the Programming-by-Demonstration (PbD) approach and on the other we approach a social interaction of the robot by the recognition of human internal states from multimodal biosignals.
|M2||Modelling Scenes and Situations||Prof. Dr.-Ing. Jürgen Beyerer||Show more|
|M3||Motion and Action Models||Prof. Dr.-Ing. Tanja Schultz, Prof. Dr.-Phil. Hermann Schwameder, Dr.-Ing. Annika Wörner||Show more|
The goal of this project is the analysis of complex human whole-body movements according to their functions and goals to transfer them into the working area of the robot. The main applications are the modeling and recognition of human movements as well as the generation of humanoid motion trajectories for the robot.
|M5||Modeling of Task-oriented Full Body Movements||Prof. Dr.-Phil. Hermann Schwameder, Prof. Dr.-Ing. Wolfgang Seemann||Show more|
The project M5 is destined to develop anthropometric models including detailed skeletal models with anthropomorphic joints, muscle and wobbling mass models. Biomechanical analysis of recorded motion and synthesis of task-oriented upper body motion in terms of optimality are used to investigate human motor control of manipulation tasks. In order to understand bipedal locomotion methods from nonlinear dynamics are applied.
|P1||Multimodal Interaction and Dialog||Prof. Dr. rer. nat. Alex Waibel||Show more|
Workpackage P1 provides components for dialog processing and speech recognition. Within the dialog component multimodal input of the recognition components is interpreted in isolation as well as with respect to the context and information is requested. The dialog constitutes an interface through which the robot can learn interactively and acquire new knowledge. Among its main tasks is also the robust recognition of continuous speech given the special constraints of the demonstrator.
|P2||Interactive Multimodal Exploration||Prof. Dr.-Ing. Jürgen Beyerer, Prof. Dr.-Ing. Kristian Kroschel, Prof. Dr.-Ing. Rainer Stiefelhagen||Show more|
This subproject aims to allow the robot to explore an unknown environment using different modalities (i.e. vision or acoustics) and under the supervision of a human. In order to help the robot to learn new objects and recognize them independently at a later time, it is necessary for the robot to percept people and their focus of interest.
|P4||Situation and Intention Recognition||Prof. Dr.-Ing. Uwe D. Hanebeck, Prof. Dr.-Ing. Rainer Stiefelhagen||Show more|
This subproject is concerned with the automatic recognition of situations and user intentions. To this end, view-based situation assessment, articulated body-tracking as well as additional domain knowledge are used. Subsequently, the user's intentions are estimated with respect to the predominant situation. The employed models shall adapt to changing environments.
|R1||System Integration, Sensorimotor Coordination and Task Execution||Dr.-Ing. Tamim Asfour, Prof. Dr.-Ing. Rüdiger Dillmann, Prof. Dr.-Ing. Jörg Henkel||Show more|
The subproject R1 is the system integration project within the CRC 588. The goal of the research activities is the integration of all developed machatronical components and functionalities into a complete humanoid robot system able to act and interact in the real world. R1 is responsible for the specification of the demonstrators, software and hardware system architecture as well as realization of integrated demonstrations. In addition, R1 deals with the research topics related to vision-based grasping, motion planning, motion coordination as well as new low-power computer architectures for humanoid robots.
|R2||Whole-Body Control||Dr.-Ing. Helge-Björn Kuntze, Dr.-Ing. Ralf Mikut||Show more|
The subproject R2 is responsible for the investigation of advanced control concepts for the multi-sensory interaction of the robot with its dynamic weakly structured environment. A discrete-continuous control architecture has been developed which will enable a perceptually controlled whole-body motion including bipedal locomotion. Currently existing neuro-fuzzy algorithms are amended by biologically inspired concepts.
|R3||Touching and Grasping||Prof. Dr.-Ing. habil. Georg Bretthauer, Prof. Dr.-Ing. Heinz Wörn||Show more|
The subproject R3 examines lightweight and flexible hands for a humanoid robot. The contributions include a position and force-torque control and the optimization of the gripping patterns based on grasp pattern database. At a higher level, the subproject considers the extension of exploration and grasping abilities, the study of exploration strategies and experience-based grasping.
|R6||Body Balancing for Humanoid Robots||Dr.-Ing. Tamim Asfour, Dr.-Ing. Stefan Schulz||Show more|
The subproject R6 investigates research questions of body balancing and the coordination between bipedal locomotion and manipulation. R6 is developing the first legs for ARMAR-IV, the next robot generation of the CRC 588 and investigates methods for the generation of stable walking patterns based on human motion data. In addition, the subproject develops new fluidic actuators for the realizations of legs for humanoid robots.
|Z1||Central Tasks||Prof. Dr.-Ing. Rüdiger Dillmann, Dipl.-Inform. Nikolaus Vahrenkamp||Show more|
|Z2||Production Preparation and Assembling of the Demonstrators||Dr.-Ing. Wolfgang Burger||Show more|
Goals of project Z2 are the integrated development and manufacturing of reproducible humanoid robot systems. This includes finding ideas, developing new innovative concepts as well as implementation in complete 3D CAD models. These models are the basis for production of parts and assembly of the respective (sub-) systems in the in-house robot manufacture of the collaborative research center 588.
|TPÖ||Public Science in Collaborative Research Centers: Inside Science||Prof. Dr. Caroline Y. Robertson||Show more|
The project InsideScience (TPÖ) experiments with new video and online means to explain complex aspects of research to and to interact with a broad public. Topics in the funding period are the Computational Particle Physics and Humanoid Robots. Scientists from both research fields cooperate with journalists and directors from the ZAK │Center for Cultural and General Studies and the Public Relations and Marketing Department (PKM) at KIT in producing videos that make research comprehendible. The theoretical basis provides the concept "public science" in the age of New Media. To achieve a balanced co-evolution of science and society, “public science” has to help people to understand research and to participate in social discourses. See more.