From October 9-14, 2016 we participated on the International Conference on Intelligent Robots and Systems held in Daejeon, Korea where we presented our recent research “Estimating Perturbations from Experience using Neural Networks and Information Transfer”. Furthermore, we participated at the “State Estimation and Terrain Perception” workshop where our poster about “A Comparing View on Robot Odometry in Underground Mining” received the best poster award. You can download the corresponding papers in the publication section.
From the 30th of June to the 3rd of July, the RoboCup celebrated its 20th anniversary at the exhibition Centre Leipzig, Germany. As member of the Mining-RoX team I used the opportunity to present actual research results to the visitors by bringing them in touch with our robots. More details can be found here.
From september 16-21 may 3, 2016 we participated on the International Conference on Robotics and Automation in Stockholm, Sweden. Furthermore, we presented our recent research “Experience-based Torque Estimation for an Industrial Robot”. You can download the corresponding paper in the publication section.
Here is a brief description:
Robotic manipulation tasks often require the control of forces and torques exerted on external objects. This paper presents a machine learning approach for estimating forces when no force sensors are present on the robot platform. In the training phase, the robot executes the desired manipulation tasks under controlled conditions with systematically varied parameter sets. All internal sensor data, in the presented case from more than 100 sensors, as well as the force exerted by the robot are recorded. Using Transfer Entropy, a statistical model is learned that identifies the subset of sensors relevant for torque estimation in the given task. At runtime, the model is used to accurately estimate the torques exerted during manipulations of the demonstrated kind. The feasibility of the approach is shown in a setting where a robotic manipulator operates a torque wrench to fasten a screw nut. Torque estimates with an accuracy of well below +-1Nm are achieved. A strength of the presented model is that no prior knowledge of the robot’s kinematics, mass distribution or sensor instrumentation is required.
From september 28 until october 3, 2015 we participated on the International Conference on Intelligent Robots and Systems in Hamburg, Germany. Furthermore, we presented our recent research “Learning to Estimate Forces for Safe Tool Usage” at the workshop “Safety for Human-Robot Interaction in Industrial Settings”. You can download the corresponding paper in the publication section.
From november 18-20, 2014 we participated on the International Conference on Humanoid Robots in Madrid, Spain. Furthermore, we presented our recent research utilizing “Transfer Entropy for Feature Extraction in Physical Human-Robot Interaction: Detecting Perturbations from Low-Cost Sensors”. A video showing the results of the experiments can be found here: YouTube. You can download the corresponding paper in the publication section.
From august 25-29, 2014 we participated on the International Symposium on Robot and Human Interactive Communication in Edinburgh, Scotland. Furthermore, we presented current results of using Dynamic Mode Decomposition for machine learning in human robot interaction.
Here is a brief description:
In many settings, e.g. physical human-robot interaction, robotic behavior must be made robust against more or less spontaneous application of external forces. Typically, this problem is tackled by means of special purpose force sensors which are, however, not available on many robotic platforms. In contrast, we propose a machine learning approach suitable for more common, although often noisy sensors. This machine learning approach makes use of Dynamic Mode Decomposition (DMD) which is able to extract the dynamics of a nonlinear system. It is therefore well suited to separate noise from regular oscillations in sensor readings during cyclic robot movements under different behavior configurations. We demonstrate the feasibility of our approach with an example where physical forces are exerted on a humanoid robot during walking. In a training phase, a snapshot based DMD model for behavior specific parameter configurations is learned. During task execution the robot must detect and estimate the external forces exerted by a human interaction partner. We compare the DMD-based approach to other interpolation schemes and show that the former outperforms the latter particularly in the presence of sensor noise. We conclude that DMD which has so far been mostly used in other fields of science, particularly fluid mechanics, is also a highly promising method for robotics. You can download the corresponding paper in the publication section.
From november 18-22, 2013 we participated, amongst about 40 other computer scientists from all over the world, on the International Autumn School 2013 on „Human – Robot – Interaction“ in Dresden, Germany. There were many amazing talks held by popular researchers as Dr.-Ing. Sami Haddadin and Prof. Jan Peters about the field of human robot interaction and machine learning. For us, the goal of the autumn school was to get an overview of the current state of the art and find some possible future directions for our research.
From october 15-17, 2013 we participated on the International Conference on Humanoid Robots in Atlanta, USA. Furthermore, we presented current results of our guidance controlled human-robot-interaction approach for a cooperative human-robot transportation task. A video showing the results of the experiments can be found here: YouTube. You can download the corresponding paper in the publication section.
The Narrator Project of the Humanoid Robotics Group Freiberg won the BuddyPaddy competition 2013 at the Symposium on Emergent Trends in Artificial Intelligence & Robotics . Visit http://www.buddypaddy.com/ for a video of the awards show.
From May 6-10, 2013 we participated on the International Conference on Robotics and Automation in Karlsruhe, Germany. Furthermore, we presented current results of our behavior adaptation approach in cooperative human-robot transportation in the “Semantics, Identification and Control of Robot-Human-Environment Interaction” workshop.