Julian Schill

ARMAR-4: A 63 DOF torque controlled humanoid robot

We present the mechatronic design of the next generation of our humanoid robots, the humanoid robot ARMAR-4, a full body torque controlled humanoid robot with 63 active degrees of freedom, 63 actuators, 214 sensors, 76 microcontroller for low-level control, 3 PCs for perception, high-level control and balancing, a weight of 70 kg including batteries and total height of 170 cm. In designing the robot we follow an integrated approach towards the implementation of high performance humanoid robot systems, able to act and interact in the real world using only on-board sensors and computation power. Special attention was paid to the realization of advanced bimanual manipulation and locomotion capabilities. The paper presents the design concept of the robot and its mechatronic realization.

Haptic object recognition for multi-fingered robot hands

In this paper, we present an approach for haptic object recognition and its evaluation on multi-fingered robot hands. The recognition approach is based on extracting key features of tactile and kinesthetic data from multiple palpations using a clustering algorithm. A multi-sensory object representation is built by fusion of tactile and kinesthetic features. We evaluated our approach on three robot hands and compared the recognition performance using object sets consisting of daily household objects. Experimental results using the five-fingered hand of the humanoid robot ARMAR, the three-fingered Schunk Dexterous Hand 2 and a parallel Gripper are performed. The results show that the proposed approach generalizes to different robot hands.

Force position control for a pneumatic anthropomorphic hand

Robot hands based on fluidic actuators are a promising technology for humanoid robots due to their compact size and excellent power-weight-ratio. Yet, such actuators are difficult to control due to the inherent nonlinearities of pneumatic systems. In this paper we present a control approach based on a simplified model of the fluidic actuator providing force and position control and further fingertip contact detection. We have implemented the method on the microcontroller of the human hand sized FRH-4 robot hand with 8 DoF and present results of several experiments, including system response and force controlled operation.

Learn to wipe: A case study of structural bootstrapping from sensorimotor experience

In this paper, we address the question of generative knowledge construction from sensorimotor experience, which is acquired by exploration. We show how actions and their effects on objects, together with perceptual representations of the objects, are used to build generative models which then can be used in internal simulation to predict the outcome of actions. Specifically, the paper presents an experiential cycle for learning association between object properties (softness and height) and action parameters for the wiping task and building generative models from sensorimotor experience resulting from wiping experiments. Object and action are linked to the observed effect to generate training data for learning a non-parametric continuous model using Support Vector Regression. In subsequent iterations, this model is grounded and used to make predictions on the expected effects for novel objects which can be used to constrain the parameter exploration. The cycle and skills have been implemented on the humanoid platform ARMAR-IIIb. Experiments with set of wiping objects differing in softness and height demonstrate efficient learning and adaptation behavior of action of wiping.

Discovery, segmentation and reactive grasping of unknown objects

Learning the visual appearance and physical properties of unknown objects is an important capability for humanoid robots that are supposed to be working in an open environment. We present an approach that enables a robot to discover new, unknown objects, segment them from the background and grasp them. This gives the robot full control over the object and allows its further multimodal exploration.

Eine fluidisch angetriebene anthropomorphe Roboterhand

Zusammenfassung In diesem Artikel wird eine fluidisch angetriebene Roboterhand vorgestellt, deren flexible Aktuatoren adaptives Greifen sowie die weiche Handhabung von Gegenständen ermöglichen. Der mechanische Aufbau vereinigt die anthropomorphe Erscheinung mit einer präzisen DreiPunkt-Kinematik eines Greifers. Es werden Designmerkmale, das Antriebskonzept, die Steuerungskomponenten, technische Eigenschaften und ein Anwendungsbeispiel für die Service-Robotik beschrieben. Abstract A fluidic driven robotic hand is presented in this article. Flexible fluidic actuators of high power-to-weight ratio are used for the actuation and allow for adaptive grasping, compliant handling and stable holding of objects during manipulation. The construction of the robotic hand combines a human-like appearance with a preciseness of a robotic gripper. This study represents a design concept, an actuation principle, set of components, technical characteristics of a fluidic hand and its application.