Refine
Has Fulltext
- no (30)
Year of publication
Document Type
- Conference Proceeding (14)
- Article (8)
- Other (6)
- Part of a Book (1)
- Doctoral Thesis (1)
Is part of the Bibliography
- no (30)
Durch die gemeinsame Bearbeitung eines Bauteils mit kooperierenden Industrierobotern kann in der Montageautomatisierung eine deutliche Taktzeitverkürzung erreicht werden. Die Programmerstellung für die unabhängigen Handhabungseinrichtungen gestaltet sich im Praxisbetrieb jedoch häufig als sehr zeitaufwändig und fehleranfällig. Aus diesem Grund wurde in einem Kooperationsprojekt des Lehrstuhl FA PS mit dem Systemhersteller Reis Robotics ein Lösungsansatz entwickelt, mit dem automatisch zeitlich und räumlich koordinierte Roboterprogramme erstellt werden können.
RRT* is a practical and efficient incremental sampling-based motion planning algorithm. However, its searching ability is quite inefficient in some cases, due to relying on uniform random sampling like other RRT-based algorithms without taking the environment information and prior knowledge into account, which particularly leads to many sampling failures or generation of useless nodes in complex environments. In this paper, we propose an extension of RRT* based on a self-learning strategy and a hybrid-biased sampling scheme to improve the planning efficiency. By taking advantage of the prior knowledge accumulation and cost estimation, the searching tree has higher probability and success rate to extend in difficult areas. We also demonstrate the performance of our algorithm by building some simulation environments for our mobile robot and conclude with the results compared with RRT*.
A significant challenge when using industrial robots as flexible handling devices - for example in the area of creating miniaturized conductive patterns on three dimensional molded interconnect devices - is their relatively low absolute accuracy, which is caused by inaccuracies in the robots kinematic chain. This characteristic has to be considered, especially if complex and precise 3D-movements are needed. One approach to overcome this problem investigated at FAPS is the usage of a specially designed robot control system with parallel data processing for real-time correction value determination and usage of these values for remote control of a robot to improve its accuracy.
The developed control system utilizes existing network technologies and uses data of sensor systems for continuous process monitoring to derive the robots actual state in course of its movement. In case of a high speed camera the image data is acquired by the control system via a first thread and subsequently the actual robot position is calculated based on this data. At the same time, for determining the robots target state, the system continuously requests the robots nominal position from the robot controller by utilizing a parallel second thread. The information describing the robots actual state is then compared with the sensor data for calculating offset values. These offset values are subsequently filtered and a correction value is calculated. In a third parallel thread new movement instructions are generated, considering the offset values as well as the robots target trajectory. Finally, for remote control of the robot, these movement instructions are transferred to the robot controller.
For almost 20 years, dielectric elastomer actuators (DEAs) have been the subject of intense research in material science. A large number of publications describe artificial muscles based on DEAs as a promising alternative for an energy efficient, lightweight and flexible actuation architecture. DEAs could improve and extend the capabilities of robotic and prosthetic devices in terms of their dynamical performance and their eligibility for energy autarkic operation. However, up to now DEAs are not available on a large scale with reproducible characteristics nor are they yet usable on a system integration level. In this paper we present recent findings of our ongoing five-year project to qualify DEAs as feasible artificial muscles for usage in compliant robot kinematics and soft prosthetic devices. The focus of this contribution lies on a new automated production process using Aerosol Jet Printing for stacked DEAs with very thin layers for reduced driving voltages and improved mechanical characteristics resulting from the additive manufacturing. Secondly, a new set of lightweight power electronics based on pulse width modulation (PWM) is presented which aims at the improvement of the overall specific power of DEA driven kinematic systems.
When robots and human workers team up, safety should be ensured at all times. In order to improve safety of human–robot collaboration in hybrid manufacturing processes, many different types of compliant robotic drives, serial elastic actuators for instance, have been developed to date. However, most of them still consist of rigid mechanical components in combination with prevailing servo motor and thus bring several disadvantages such as poor power-to-weight ratio. This paper presents our solution approach for realization of biomimetic, inherently compliant artificial muscles based on dielectric elastomer actuators (DEAs). The artificial muscles based on DEAs distinguish themselves from conventional actuators through their favorable characteristics. They (a) work noiseless, (b) feature specific energy density comparable to human skeletal muscles and (c) are capable of storing or even recovering energy, and finally, (d) can adjust their geometry to meet with undefined and unstructured objects or environments. With remaining challenges overcome, the DEAs are expected to provide a significant contribution to safety of industrial robots collaborating with human workers.
This paper introduces an approach for enabling visually impaired and blind people to practice jogging activities by 3D environment perception for course detection and collision avoidance, as well as feedback generation in an intuitive manner. Besides a system concept, first prototypic realizations, that confirm the general feasibility, are presented for this domain, which has not been addressed by research until now.
Ganzheitliche Automatisierung mechatronischer Systeme in der Medizin am Beispiel Strahlentherapie
(2012)
Ziel dieser Arbeit war es, technologische und methodische Optimierungsansätze zu entwickeln und zu untersuchen, mittels derer effiziente, sichere und qualitativ hochwertige mechatronische Behandlungssysteme realisiert werden können. Das entwickelte hochflexible und universell einsetzbare Strahlentherapiegerät eröffnet dabei ein unerreicht breites Applikationsspektrum und lässt sich aufgrund vielfältiger neuartiger Assistenzfunktionen sehr gut in die klinikspezifischen Workflows integrieren. In dieser Arbeit konnte gezeigt werden, dass durch die integrierte Entwicklung von Behandlungssystemen unter Einbezug der mechatronischen Positioniergeräte sowie bedarfsgerechter automatisierungstechnischer Assistenzsysteme erhebliche Effizienz- und Qualitätssteigerungen in der Medizin erzielt werden können.
Patient handling robots are increasingly employed to enable a flexible positioning of the patient for diagnostic and therapeutic purposes. Due to the greatly differing robot loads because of varying patient weights especially serial kinematics are no longer able to keep up with the constantly increasing medical demands for positioning accuracy. In this chapter a low-cost measuring system for a permanent integration into the workspace of a patient handling robot is introduced which can measure the pose of a patient couch with high accuracy. This enables closed-loop control of the patient couch. In tests on a robot system an average positioning error of 0.12 mm was achieved for a diversity of medically relevant poses.
This paper presents the LinKin System, a fundamentally new robotic radiation therapy device with six degrees of freedom. Because of its linear kinematics system design it has substantial advantages over common therapy devices. The highly flexible and universally applicable LinKin System opens a unique broad application range. Furthermore, due to its novel assistant functions it can be well-integrated into the clinic specific workflows.
A radiotherapy installation is described, with a treatment head which is arranged on a treatment head support and which has an exit port for a beam generated in a beam generator, and with a patient table. The treatment head support is guided on a first rectilinear guide, which is guided on a second rectilinear guide arranged perpendicular to the first rectilinear guide, in order to move the treatment head relative to the patient table during the radiotherapy.
Patient handling robots are employed more and more frequently to enable a flexible positioning of the patient for diagnostic and therapeutic purposes. Due to the greatly differing robot loads because of varying patient weights especially serial kinematics are no longer able to keep up with the constantly increasing medical demands for positioning accuracy. In this paper a camera-based measuring system is introduced which can measure the pose of a patient couch mounted on a patient handling robot with high accuracy. Through the use of cost-effective hardware components the stereoscopic system is suitable for a permanent integration into the workspace of the robot. This enables online pose control of the patient couch. In tests on a robot system a maximum positioning error of 0.25 mm was achieved for a diversity of medically relevant poses.