Papers by Rudiger Dillmann
Extraction, evaluation, selection and classification of motion features for human activity recognition
... Sebastian Brännström and Henrik I. Christensen Centre for Autonomous Systems Royal Institute ... more ... Sebastian Brännström and Henrik I. Christensen Centre for Autonomous Systems Royal Institute of Technology (KTH) SE - 100 44 Stockholm, Sweden Email: teknolog@gmail.com, hic@nada.kth.se ... [5] [6] and Nascimento et al. [7], which rely on large area camera images. ...
An autonomous ice-cream serving robot
2011 IEEE International Conference on Robotics and Automation, 2011
An autonomous ice cream serving robot is pre- sented in this video. The video was filmed during t... more An autonomous ice cream serving robot is pre- sented in this video. The video was filmed during the Automatica 2010 trade fair in Munich. Within four days, approximately 250 scoops of different kinds of ice cream have been served to visitors during the fair. Using the KUKA light-weighted robotic arms and the DLR/HIT robotic hands, two scientific aspects are shown
Grasp planning: Find the contact points
2007 IEEE International Conference on Robotics and Biomimetics (ROBIO), 2007
ABSTRACT
2008 IEEE/ASME International Conference on Advanced Intelligent Mechatronics, 2008
Automatic grasp planning systems are very important for service robots, which compute what forces... more Automatic grasp planning systems are very important for service robots, which compute what forces should be exerted onto the object and how those forces can be applied by robotic hands. In this paper, a highly integrated grasp planning system is introduced. Initial grasp is computed in the grasp simulator GraspIt! combining hand preshapes and automatically generated approach directions. With fixed relative position and orientation between the robotic hand and object as by the initial grasp, all the contact points between the fingers and the object are efficiently found. A search process tries to improve the grasp quality by moving the fingers to its neighbored joint positions, and uses the corresponding contact points to the joint position to evaluate the grasp quality, until local maximum grasp quality is reached. Optimal forces for the found grasp is computed as a linear inequalities matrix problem, which are exerted onto the object using torque based finger impedance control during execution. Experiments on Schunk Anthropomorphic Hand with 13 degrees of freedom show that, using the introduced grasp planning system, the object can be grasped solidly with shift errors of only some millimeters.
This report describes the work performed at University of Karlsruhe concerning Human Activity Rec... more This report describes the work performed at University of Karlsruhe concerning Human Activity Recognition in the context of the EU project Cogniron. Within another part of the project, human motion and body configuration is observed and tracked. This intrinsic information is used together with extrinsic features to recognize and classify human activities. Recognition is done using a 3-step approach: (1) From human body motion, a feature set is extracted. (2) These features are evaluated regarding their relevance for each activity, and (3) for each activity, a FFNN classifier computes a likelihood based on manually segmented training data sets. This report gives a state of the art, and then describes in detail the depicted activity recognition approach. Experiments and results are given which show that this method works with over 100 features and more than 10 activities.
Online planning of action sequences for a two-arm manipulator system
Proceedings 1992 IEEE International Conference on Robotics and Automation
... the origi-nal assembly and removes the obstacle by generating a ,,pick-and-place ... CM; Requ... more ... the origi-nal assembly and removes the obstacle by generating a ,,pick-and-place ... CM; Requicha, AAG; Voelcker, HB: Geo-metric modelling systems for mechanical design and manufacturing ... Meier W.; Graf, J.: A two-arm robot system based on trajectory optimization and hybrid ...
A Petri net based control architecture for a multi-robot system
International Symposium on Intelligent Control, Sep 25, 1989
An effort is made to show how Petri nets can be applied to control an advanced multirobot system.... more An effort is made to show how Petri nets can be applied to control an advanced multirobot system. The action plan for the robot system, i.e. the description of the task the robot has to perform, is obtained by using a special class of condition/event Petri nets. With the aid of a distributed multiprocessor system, the robot system is able to execute net actions in parallel if necessary. The use of the programming language Prolog for both the specification and the interpretation of the Petri net supports rapid prototyping and offers a powerful framework for integrating rule-based diagnosis and error recovery components into the robot control system.<<ETX>>
Sensorsimulation-Based Interactive Telerobot Control System
ABSTRACT
Interactive Robot Control System for Teleoperation
ABSTRACT
Integration of Sensorsimulation for Telerobotic and Telepresence
ABSTRACT
Grid analysis for mesh-independent and stable CFD simulations in the aorta
ABSTRACT
Rule-based tracking of multiple lanes using particle filters
2006 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems, 2006
Abstract-Tracking of lanes is essential for intelligent ve-hicles in order to drive autonomously.... more Abstract-Tracking of lanes is essential for intelligent ve-hicles in order to drive autonomously. A system is presented which allows tracking of multiple lanes. The system is based on a clear modelling of a lane and the parameterset of each lane is estimated using a ...
Fusing image features and navigation system data for augmenting guiding information displays
2006 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems, 2006
Abstract-Today's navigation systems provide their guiding information using speech and a dis... more Abstract-Today's navigation systems provide their guiding information using speech and a display with arrows or a map together with the car's position. Recent advances in computer vision offer the ability to extract more detailed information about the environment and to combine it ...
2006 IEEE International Conference on Multisensor Fusion and Integration for Intelligent Systems, 2006
In this paper, we present a new approach for fusion of different measurements and sensors for 3D ... more In this paper, we present a new approach for fusion of different measurements and sensors for 3D model based tracking. The underlying model of the tracked body is defined geometrically with generalized cylinders, which can hierarchically be connected by different kinds of joints. This results in an articulated body model with constrained kinematic degrees of freedom. The fusion approach incorporates this model knowledge together with the measurements, and tracks the target body iteratively with an extended Iterative Closest Point approach. The resulting tracking system named VooDoo is used to track humans in a Human-Robot Interaction (HRI) context. We only rely on sensors on board the robot, i.e. a color camera, a 3d time-of-flight camera and a laser range finder. The system runs in realtime (∼ 20Hz) and is able to robustly track a human in the vicinity of the robot. The pose and trajectory of the human interaction partner can then be used for haptic interaction like handovers , and for activity and gesture recognition.
Computational blood flow and vessel wall modeling in a CT-based thoracic aorta after stent-graft implantation
SPIE Proceedings, 2010
Abnormal blood flow conditions and structural fatigue within stented vessels may lead to undesire... more Abnormal blood flow conditions and structural fatigue within stented vessels may lead to undesired failure causing death to the patient. Image-based computational modeling provides a physical and realistic insight into the patientspecific biomechanics and enables accurate predictive simulations of development, growth and failure of cardiovascular diseases as well as associated risks. Controlling the efficiency of an endovascular treatment is necessary for the evaluation of potential complications and predictions on the assessment of the pathological state. In this paper we investigate the effects of stent-graft implantation on the biomechanics in a patient-specific thoracic aortic model. The patient geometry and the implanted stent-graft are obtained from morphological data based on a CT scan performed during a controlling routine. Computational fluid dynamics (CFD) and computational structure mechanics (CSM) simulations are conducted based on the finite volume method (FVM) and on the finite element method (FEM) to compute the hemodynamics and the elastomechanics within the aortic model, respectively. Physiological data based on transient pressure and velocity profiles are used to set the necessary boundary conditions. Further, the effects of various boundary conditions and definition of contact interactions on the numerical stability of the blood flow and the vessel wall simulation results are also investigated. The quantification of the hemodynamics and the elastomechanics post endovascular intervention provides a realistic controlling of the state of the stented vessel and of the efficiency of the therapy. Consequently, computational modeling would help in evaluating individual therapies and optimal treatment strategies in the field of minimally invasive endovascular surgery.
Humanoid Robots: New Developments, 2007
This paper describes ongoing effort in developing an anthropomorphic arm/hand manipulation system... more This paper describes ongoing effort in developing an anthropomorphic arm/hand manipulation system as part of the German Humanoid Research Project SFB588. This project aims at building a humanoid robot that assists humans in everyday service tasks. The manipulation part of this project has to deal with a service environment and to cope with the wide variety of objects and activities encountered in dynamic unstructured environments. Here, we describe several research lines aimed to fulfill these requirements. They consist in building anthropomorphic robotic hands and arms that integrate sensorial capabilities, developing manipulation algorithms that make use of such capabilities, and developing an architecture able to integrate the control of this subpart in the whole humanoid control system.
Planning regrasp operations for a multifingered robotic hand
2008 IEEE International Conference on Automation Science and Engineering, 2008
ABSTRACT
Humanoids 2008 - 8th IEEE-RAS International Conference on Humanoid Robots, 2008
Haptic exploration of unknown objects is of great importance for acquiring multi-modal object rep... more Haptic exploration of unknown objects is of great importance for acquiring multi-modal object representations, which enable a humanoid robot to autonomously execute grasping and manipulation tasks. In this paper we present a tactile exploration strategy to guide an anthropomorphic five-finger hand along the surface of previously unknown objects and build a 3D object representation based on acquired tactile point clouds. The proposed strategy makes use of the dynamic potential field approach suggested in the context of mobile robot navigation. To demonstrate the capabilities of this strategy, we conduct experiments in a detailed physics simulation using a model of the five-finger hand. Exploration results of several test objects are given.
2012 4th IEEE RAS & EMBS International Conference on Biomedical Robotics and Biomechatronics (BioRob), 2012
Grasp stability estimation with complex robots in environments with uncertainty is a major resear... more Grasp stability estimation with complex robots in environments with uncertainty is a major research challenge. Analytical measures such as force closure based grasp quality metrics are often impractical because tactile sensors are unable to measure contacts accurately enough especially in soft contact cases. Recently, an alternative approach of learning the stability based on examples has been proposed. Current approaches of stability learning analyze the tactile sensor readings only at the end of the grasp attempt, which makes them somewhat time consuming, because the grasp can be stable already earlier. In this paper, we propose an approach for grasp stability learning, which estimates the stability continuously during the grasp attempt. The approach is based on temporal filtering of a support vector machine classifier output. Experimental evaluation is performed on an anthropomorphic ARMAR-IIIb. The results demonstrate that the continuous estimation provides equal performance to the earlier approaches while reducing the time to reach a stable grasp significantly. Moreover, the results demonstrate for the first time that the learning based stability estimation can be used with a flexible, pneumatically actuated hand, in contrast to the rigid hands used in earlier works.
Uploads
Papers by Rudiger Dillmann