Dielectric elastomer sensors (DES) are compliant systems that, allow the detection of geometric changes caused by external forces. However, a simple DES provides only the amplitude of a deformation. In order to fully describe a tensile force vector, it is necessary to characterize it by its length, direction, and origin. Therefore, there are many approaches to improve the information that can be obtained from the sensor, such as arranging DES patterns or modifying the impact of the applied mechanical force by using external components. To improve and parallelize the information provided by the mechanical deformation of the system, we aim to combine the operation principle of structured resistive strain sensors with the capacitive properties of DES. By structuring the electrodes with different patterns and combining two dissimilar electrodes in one setup, it is possible to detect the length and direction of a force vector with one sensing element. To study the patterning and its effect on resistive and capacitive signals, we use an aerosol jet printing technique that allows selective deposition of a conductive ink. The printed lines show a higher resistance increase when forced perpendicular to the pattern and vice versa, which allows to distinguish the direction and the type of applied force. In this study, the influence of printing parameters and signal interpretation for different forces are investigated. The results show that the combination of resistive and capacitive signals allows discrimination between different motions.

Herkömmliche Roboter sind aufgrund der Konfiguration rigider Gelenke und empfindlicher Getriebe anfällig für Stoßbelastungen. In diesem Beitrag wird gezeigt, dass ein Robotergelenk, basierend auf einer neuartigen Topologie mit drei aktiven und drei passiven Seilen, Stoßbelastungen bewältigen kann, die bei Bearbeitungsaufgaben wie dem Hämmern auftreten können.   Conventional robots are vulnerable to high impact loads due to the configuration of rigid joints and fragile gears. This paper shows that a robotic joint based on a novel topology with three active and three passive cables can cope with impact loads as they occur in machining tasks such as hammering.

Beyond conventional automated tasks, autonomous robot capabilities aside human cognitive skills are gaining importance in industrial applications. Although machine learning is a major enabler of autonomous robots, system adaptation remains challenging and time-consuming. The objective of this research work is to propose and evaluate an augmented virtuality-based input demonstration refinement method improving hybrid manipulation learning for industrial bin picking. To this end, deep reinforcement and imitation learning are combined to shorten required adaptation timespans to new components and changing scenarios. The method covers initial learning and dataset tuning during ramp-up as well as fault intervention and dataset refinement. For evaluation standard industrial components and systems serve within a real-world experimental bin picking setup utilizing an articulated robot. As part of the quantitative evaluation, the method is benchmarked against conventional learning methods. As a result, required annotation efforts for successful object grasping are reduced. Thereby, final grasping success rates are increased. Implementation samples are available on: https://github.com/FAU-FAPS/hybrid_manipulationlearning_unity3dros