6 results
Filters
Settings
Now showing1 - 6 of 6
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Generation of synthetic AI training data for robotic grasp-candidate identification and evaluation in intralogistics bin-picking scenariosRobotic bin picking remains a main challenge for the wide enablement of industrial robotic tasks. While AI-enabled picking approaches are encouraging they repeatedly face the problem of data availability. The scope of this paper is to present a method that combines analytical grasp research with the field of synthetic data creation to generate individual training data for use-cases in intralogistics transportation scenarios. Special attention is given to systematic grasp finding for new objects and unknown geometries in transportation bins and to match the generated data to a real two-finger parallel gripper. The presented approach includes a grasping simulation in Pybullet to investigate the general tangibility of objects under uncertainty and combines these findings with a previously reported virtual scene generator in Blender, which generates AI-images of fully packed transport boxes, including depth maps and necessary annotations. This paper, therefore, contributes a synthesizing and cross-topic approach that combines different facets of bin-picking research such as geometric analysis, determination of tangibility of objects, grasping under uncertainty, finding grasps in dynamic and restricted bin-environments, and automation of synthetic data generation. The approach is utilized to generate synthetic grasp training data and to train a grasp-generating convolutional neural network (GG-CNN) and demonstrated on real-world objects.Publicationtype: Conference PaperScopus© Citations 1 84 122 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Assessing visual identification challenges for unmarked and similar aircraft componentsHighest demands for complete traceability and quality control of each component, require thorough identification of each produced, replaced, and (dis-)assembled aircraft component. As many production and MRO-processes for modern aircraft remain to be carried out manually, this poses a great challenge. Many small components either do not feature a Part Number or in MRO-processes their Part Number is occluded or not readable due to dirt and wear. Considering unmarked components with a high resemblance to one another and few characteristics, e.g. standard parts such as bushings and pipes, manual identification is an error-prone task. Avoiding errors through digitalized procedures has the potential to significantly reduce error rates and costs for a typical manual dual control. However, automated identification of components has to overcome the high classification complexity that originates in the manifold of aircraft components and is additionally increased by individualistic MRO modifications for specific aircraft. This work presents a methodological approach to reveal possible challenges for identification procedures and gives special focus to the assessment of similarities between components. Two similarity metrics are introduced that are calculated either through feature-based analysis or through 3D-shape similarity assessment. The methodology is demonstrated with two to this date unsolved Use-Cases that represent different challenges of visual identification systems for similar and unmarked components.Publicationtype: Conference PaperScopus© Citations 1 55 58 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Die-less forming of fiber-reinforced plastic compositesFiber-reinforced plastics (FRP) are increasingly popular in light weight applications such as aircraft manufacturing. However, most production processes of thin-walled FRP parts to date involve the use of expensive forming tools. This especially hinders cost-effective production of small series as well as individual parts and prototypes. In this paper, we develop new possible alternatives of highly automated and die-less production processes based on a short review of current approaches on flexible thin-walled FRP production. All proposed processes involve robot guided standard tools, similar to incremental sheet metal forming, for local forming of the base materials. These include woven glass fiber fabrics which are locally impregnated with thermoset resin and cured using UV-light, woven commingled yarns made out of glass fibers and thermoplastic fibers which are locally heated and pressed, as well as pre-consolidated thermoplastic organo sheets which require selective heating for forming. General applicability of the processes is investigated and validated in practical experiments.Publicationtype: Conference PaperScopus© Citations 3 160 153 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Development process for information security concepts in IIoT-based manufacturingDigital technologies are increasingly utilized by manufacturers to make processes more transparent, efficient and networked. Novel utilization elicits the challenge of preventing deployed information technology from compromising processual security. The digital enabling of formerly analog operation technology, the extensive use of information technology connectivity like MQTT, TCP/IP, Wi-Fi, and the deployment of IoT edge computing platforms create an application scenario for the Industrial Internet of Things (IIoT), which also introduces the associated vulnerabilities, which have been extensively exploited in the past. This paper introduces a development process for information security concepts designed for production scenarios based on the IIoT. This concept is then applied using an illustrative use case from aircraft production. The main contents of the development process include: Formulation of reasonable assumptions, system modelling, threat analysis including risk assessment, recommendation of countermeasures, reassessment after incorporating countermeasures. Specifically, a Data Flow Diagram as the model is developed, and a “risk first” variation of the STRIDE methodology is applied to identify threats and prioritize them. The aforementioned state-of-the-art methodologies are adjusted to our cyber-physical use case in the IIoT. The resulting concept aims to enable manufacturing processes to be digitized as sought. The adjustments to the methodologies are independent from our use case and may be suitable to a broad field of scenarios in the IIoT.Publicationtype: Conference PaperScopus© Citations 1 117 148 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Processing strategies for dieless forming of fiber-reinforced plastic composites(Multidisciplinary Digital Publishing Institute (MDPI), 2023-03); ; The demand for lightweight materials, such as fiber-reinforced plastics (FRP), is constantly growing. However, current FRP production mostly relies on expensive molds representing the final part geometry, which is not economical for prototyping or highly individualized products, such as in the medical or sporting goods sector. Therefore, inspired by incremental sheet metal forming, we conduct a systematic functional analysis on new processing methods for shaping woven FRP without the use of molds. Considering different material combinations, such as dry fabric with thermoset resin, thermoset prepreg, thermoplastic commingled yarn weave and organo sheets, we propose potential technical implementations of novel dieless forming techniques, making use of simple robot-guided standard tools, such as hemispherical tool tips or rollers. Feasibility of selected approaches is investigated in basic practical experiments with handheld tools. Results show that the main challenge of dieless local forming, the conservation of already formed shapes while allowing drapability of remaining areas, is best fulfilled by local impregnation, consolidation and solidification of commingled yarn fabric, as well as concurrent forming of prepreg and metal wire mesh support material. Further research is proposed to improve part quality.Publicationtype: Journal ArticleScopus© Citations 3 255 154 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Access status: with files , Deep anomaly detection for endoscopic inspection of cast iron partsDetecting anomalies in image data plays a key role in automated industrial quality control. For this purpose, machine learning methods have proven useful for image processing tasks. However, supervised machine learning methods are highly dependent on the data with which they have been trained. In industrial environments data of defective samples are rare. In addition, the available data are often biased towards specific types, shapes, sizes, and locations of defects. On the contrary, one-class classification (OCC) methods can solely be trained with normal data which are usually easy to obtain in large quantities. In this work we evaluate the applicability of advanced OCC methods for an industrial inspection task. Convolutional Autoencoders and Generative Adversarial Networks are applied and compared with Convolutional Neural Networks. As an industrial use case we investigate the endoscopic inspection of cast iron parts. For the use case a dataset was created. Results show that both GAN and autoencoder-based OCC methods are suitable for detecting defective images in our industrial use case and perform on par with supervised learning methods when few data are available.Publicationtype: Conference Paper106 125