检索结果-内蒙古大学图书馆

First International Conference on Transdisciplinary AI (TransAI)

作者： Ralf Gulde Khanh Nguyen Marc Tuscher Oliver Riedel Alexander Verl Institute of Control Engineering of Machine Tools and Manufacturing Units University of Stuttgart Stuttgart Germany

Reinforcement Learning (RL) has gained popularity for developing intelligent robots, but challenges such as sample inefficiency and lack of generalization persist. The choice of observation space significantly influences RL algorithms' sample efficiency in robotics. While end-to-end learning has been emphasized, it increases complexity and inefficiency as the agent must re-learn forward and inverse kinematics. To address these issues, we propose a straightforward approach that utilizes readily available control techniques, such as forward and inverse kinematics, to capitalize on domain knowledge. Our approach involves enhancing the observation space with task-space features and utilizing task-space inverse kinematics. Our contributions include a proposal for mathematical formulation and a framework for RL algorithms in robotics.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Topology Matching of Branched Deformable Linear Objects

Topology Matching of Branched Deformable Linear Objects

引用

IEEE International Conference on Robotics and Automation (ICRA)

作者： Manuel Zürn Markus Wnuk Armin Lechler Alexander Verl Institute for Control Engineering of Machine Tools and Manufacturing Units University of Stuttgart Stuttgart Germany

This paper presents a new method for correspondence estimation between a previously known topology of a branched deformable linear object and an image representation from a 3D stereo camera. Although frequently encountered in production, robotic deformable linear object manipulation still lacks reliable sensor feedback. Especially for branched deformable linear objects, such as wire harnesses, correspondence estimation is very challenging. Due to their flexible nature, they have an infinite-dimensional configuration space, such that visual appearances of the same object can vary strongly. Knowing the correspondence is vital for various applications, e.g., estimating valid grasping positions for robotic wire routing or augmented reality support for workers. Therefore, this paper presents a method for matching the topology of a branched deformable linear object to camera sensor data. Asymmetries in the wire harness design reduce the solution space by comparing the known topology of a model to the topology extracted from sensor data. The problem of finding the most likely solution to the matching problem requires features extracted from camera images. These features are used to construct a graph-based topology representation, which can then be matched to a graph-based topology representation of the known branched deformable linear object. The presented method is evaluated using multiple different non-overlapping configurations of a wire harness, showing the effectiveness of a graph-based segment matching approach.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Automatic Generation of Digital Twin Models for Simulation of Reconfigurable Robotic Fabrication Systems for Timber Prefabrication 38

Automatic Generation of Digital Twin Models for Simulation o...

引用

38th International Symposium on Automation and Robotics in Construction, ISARC 2021

作者： Kaiser, Benjamin Littfinski, Daniel Verl, Alexander Institute for Control Engineering of Machine Tools and Manufacturing Systems University of Stuttgart Germany

ISBN: (纸本)9789526952413

Timber construction and prefabrication are becoming increasingly important in the building and construction industry. The degree of automation in this area is low. Caused by the great variability of construction projects and building components, automation systems that can adapt flexibly to different construction projects are required. However, a system that fulfills these conditions does not yet exist. Down-times for reconfiguration in between projects must be short. The use of robotics for the automation of prefabrication is constantly developing. Recent publications present a novel, flexible, reconfigurable, and transportable manufacturing platform using industrial robots. A system consisting of several of these modular platforms can be flexibly adapted to the changing requirements of different construction projects and set up on-site at local timber construction companies. To ensure the manufacturability of the timber components and to minimize downtimes between projects, the entire workflow from digital design to fabrication must be considered. The co-design approach makes this possible. It breaks up the currently existing sequential process from design to fabrication and considers all sub-steps holistically. Thus, enabling the consideration of the fabrication capabilities during building design. This allows the planning of building components and system layout so that fabrication is feasible. To achieve this, the fabricability of a given system layout has to be evaluated. This requires a digital twin of the fabrication system as a simulation model. The reconfigurability of the fabrication system must be reflected by the digital twin. The fabrication tasks and the system configuration are constantly changed in search of a valid combination. Therefore, with each iteration, the digital twin must be adapted or newly created. At present, this is mostly a manual process. This makes the whole approach unfeasible. To solve this problem, this work presents an approach

关键词： Cyber-Physical System Digital Twin Simulation Software Defined manufacturing Timber Prefabrication

来源：评论

学校读者我要写书评

暂无评论

Method for the Calculation of Local Bead Volume in Multi-Axis Additive manufacturing

引用

Procedia CIRP 2023年 120卷 1244-1249页

作者： Martin Wolf Vinzent Maier Alexander Verl Institute for Control Engineering of Machine Tools and Manufacturing Units (ISW) University of Stuttgart Seidenstr. 36 70174 Stuttgart Germany

In recent history, Multi-Axis Additive manufacturing enjoys increasing attention for both industrial and research application. In comparison to traditional processes, the additional rotation axes enable parts to be printed with greater design flexibility, while reducing the need for support structures. The multi-axis motion additionally allows for optimization of structural properties of the parts. However, curved-layer printing requires significantly higher effort for planning and computation, as the layers exhibit curvature in their shape and feature local variations of the layer height and bead width. Consideration of these properties is often neglected in the development of path planning methods, although they pose a significant influence to the optical and structural quality of the printed parts. Local over- or under-extrusion can result in material accumulation or voids within the part and need to be avoided. Additionally, some Multi-Axis Additive manufacturing processes like Directed Energy Deposition are especially prone to errors in practice, as the melt-pool behavior is hard to model and predict accurately. This paper introduces a method for the precise calculation and adjustment of local bead volume, as a possible solution to these problems. The method processes existing Cartesian paths and uses a model of the bead geometry to simulate the extrusion by rasterization. The method considers printing head orientation, as well as neighborhood information of adjacent beads and builds a volumetric map of the paths. Local extrusion volume is then determined for each waypoint of the path from this volume map and provided to the printer along with the G-Code, enabling adjusted volume deposition both in advance and during the process by iterative path-adaptation. The provided implementation is able to achieve favorable results for both use-cases but proves especially beneficial for the latter.

关键词： Multi-Axis Additive manufacturing Computer-Aided manufacturing Local Volume Calculation 3D Printing

来源：评论

学校读者我要写书评

暂无评论

Co-Design of Structural Timber Components Through Automated Model Generation for manufacturing Simulation of Reconfigurable manufacturing Systems

Co-Design of Structural Timber Components Through Automated ...

引用

International Conference on Mechatronics and machine Vision in Practice (M2VIP)

作者： Benjamin Kaiser Alexander Verl Institute for Control Engineering of Machine Tools and Manufacturing Units University of Stuttgart Stuttgart Germany

manufacturing of products with high variability and small batch sizes, such as prefabricated timber building components, presents significant manufacturing challenges. In urban areas, the construction industry's relatively low productivity, coupled with rising demand for affordable housing and irregularly shaped spaces, calls for novel and innovative construction technologies. The method of computational design shows promise as a solution, allowing unique, optimized structures composed of individual components, departing from conventional modular construction. Automated prefabrication of these building components enhances construction productivity. Reconfigurable manufacturing systems offer a practical solution for addressing building component variations and changing project requirements, requiring frequent system re-configuration. The co-design process concurrently optimizes building components and manufacturing systems to ensure manufacturability and structural optimization through iterative design, validated by simulation. However, manual simulation implementation and validation for component manufacturability is time-consuming and error-prone, especially considering the numerous iterations and layout variants involved in the co-design process. This paper proposes a method for automatically generating and conducting simulations to check the manufacturability of building elements to enable co-design. The proposed method uses AutomationML as a modeling language for describing the manufacturing system and building components. We conduct several simulations with parametrically generated timber shell elements to validate the proposed method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Recurrent Neural Network for Modelling a Contractive Soft Actuator 56

Recurrent Neural Network for Modelling a Contractive Soft Ac...

引用

56th International Symposium on Robotics, ISR Europe 2023

作者： Kienzlen, Annika Zürn, Manuel Kazemi, Shahab Xu, Weiliang Cheng, Leo K. Strommel, Martin Verl, Alexander Institute for Control Engineering of Machine Tools and Manufacturing Units University of Stuttgart Germany Department of Mechanical & Mechatronics Engineering University of Auckland New Zealand Auckland Bioengineering Institute University of Auckland New Zealand Department of Electrical & Electronic Engineering University of Auckland New Zealand

ISBN: (纸本)9783800761418

Robotic stomach simulators are designed and employed to model stomach motility and gain insights into the digestive process. These simulators consist of multiple nonlinear soft actuators. Their modelling enhances the controllability and optimisation possibilities and shows their geometry variations and deformations. By leveraging a dynamic model, it becomes possible to design time-efficient model-based controllers that handle the nonlinearities of the soft robot. However, the modelling of soft actuators is challenging due to their infinite number of degrees of freedom. In this paper, we propose a novel approach that maps the dynamics of a contractive soft actuator to a neural network. The short calculation time of neural networks combined with their ability to model complex relationships make them a promising approach for controlling soft actuators. Since the contractive behaviour of the soft actuator is difficult to capture, training data can be generated using a physical simulation model. The results demonstrate that a simple recurrent neural network can model similar dynamics to a physical simulation model with significantly reduced calculation times. Since a fast and precise computer model is necessary to capture the complex dynamics of the soft actuators accurately for control purposes, the recurrent neural network is promising. © VDE VERLAG GMBH Berlin Offenbach.

关键词： contractive actuation deep learning recurrent neural network soft robotics

来源：评论

学校读者我要写书评

暂无评论

A Multicore control System Architecture as an Operating Platform for Industrial Digital Twins

引用

Procedia CIRP 2023年 118卷 294-299页

作者： Lars Klingel Karl Kübler Alexander Verl Institute for Control Engineering of Machine Tools and Manufacturing Units (ISW) University of Stuttgart Seidenstr. 36 70174 Stuttgart Germany

Digital Twins are seen as one of the core technologies and enablers of Industrie 4.0. There are various definitions of the term Digital Twin, but most agree that they are digital representations supporting the real production system throughout its lifecycle by providing models, data-integration and online-optimization. Currently, there are no implementations of Digital Twins to be used productively, which fulfill all parts of this definition. This work focuses on the simulation part of Digital Twins and the real-time requirements which arises from an online calculation. An approach that uses models of virtual commissioning, which are integrated in an industrial controller to tackle the problem of real-time calculation is presented. To enable the online simulation of complex and multi-physical models the simulation is parallelized and calculated on a multicore processor.

关键词： Industrial Digital Twin Co-simulation Partitioning Virtual Commissioning

来源：评论

学校读者我要写书评

暂无评论

Deep Learning-Based Instance Segmentation for Feature Extraction of Branched Deformable Linear Objects for Robotic Manipulation 19

Deep Learning-Based Instance Segmentation for Feature Extrac...

引用

19th IEEE International Conference on Automation Science and engineering, CASE 2023

作者： Zürn, Manuel Kienzlen, Annika Klingel, Lars Lechler, Armin Verl, Alexander Ren, Shiyi Xu, Weiliang Institute for Control Engineering of Machine Tools and Manufacturing Units University of Stuttgart Seidenstr. 36 Stuttgart70174 Germany The University of Auckland Department of Mechanical and Mechatronics Engineering 20 Symonds St. Auckland1001 New Zealand

ISBN: (纸本)9798350320695

Branched deformable linear objects, like wire harnesses, are vastly used in the electrical industry to interconnect multiple electronic components while allowing small relative motions. Their deformable nature gives them an infinite-dimensional configuration space, which is challenging for robotic manipulation and requires sensor feedback about the actual configuration. Cameras have been used before in automation attempts as sensor feedback of branched deformable linear objects. However, image processing is sensible to occlusions and unstructured environments. This paper presents a method of using deep learning-based instance segmentation networks to extract topological features like connectors, clips, branch points and overlapping points on the branched deformable linear object. The topological features originate from features that can be used for matching a stretched representation of the branched deformable linear object to its sensor representation. Our results show that the network is capable of extracting the features for different datasets with high accuracy. Furthermore, this paper evaluates the use of preliminary results in the training process to facilitate annotation by predicting the labels. For now, this paper only uses 2D images of the branched deformable linear objects. Future work will also use depth images, such that individual deformable linear object features, such as length or radius, can be extracted from the images. We anticipate the usage of wire harness automation with robotic manipulation, augmented reality support of the worker or quality control of wire harnesses. © 2023 IEEE.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Self-adaptive digital twin reference architecture to improve process quality

引用

Procedia CIRP 2023年 119卷 867-872页

作者： Ann-Kathrin Splettstößer Carsten Ellwein Andreas Wortmann Institute for Control Engineering of Machine Tools and Manufacturing Units University of Stuttgart Seidenstr. 36 D-70174 Stuttgart Germany

The aim of software-defined manufacturing is a more adaptable production through software. This requires highly adaptive, versatile cyber-physical production systems (CPPSs), which react flexibly to changes and perform crucial adjustments. Self-adaptive digital twins (DTs) enable adaptability through monitoring the behavior of a CPPS, identifying problems, and determining adjustments for the optimization of processes. DTs work with an idealized representation of the CPPS. Changes in system behavior that negatively impact the process execution, e.g., caused by tool wear or variations of environmental influences, and, thus, the quality of the resulting product, are ignored. This ignorance leads to defective products, sub-optimal product quality, and hazardous CPPS states. To mitigate this, we present an approach to incorporate quality awareness into a model-driven reference architecture of a self-adaptive DT. The devised concept automatically analyzes and optimizes system behavior. To this end, a component for quality evaluation extends the reference architecture. The DT and CPPS are connected via OPC UA. To validate the approach, a fiber laser machine is utilized. The DT can operate both on a physical and a virtual laser machine.

关键词： Digital Twin Model-Driven Development Software-Defined manufacturing Quality Awareness

来源：评论

学校读者我要写书评

暂无评论

A modular configuration and management framework for distributed real-time applications based on converged networks using TSN

引用

Procedia CIRP 2023年 118卷 38-43页

作者： Stefan Oechsle Florian Frick Armin Lechler Alexander Verl Institute for Control Engineering of Machine Tools and Manufacturing Units (ISW) University of Stuttgart Seidenstr. 36 70174 Stuttgart Germany

Industrial control systems are distributed real-time applications deployed on field and controller devices, connected by fieldbuses. Their lack of interoperability, mutually and towards IT-systems, is a major obstacle for the digitalization. Converged networks are considered as the solution across industries with Time-Sensitive Networking (TSN) as the enabling technology. To ease the adoption of TSN for industrial automation, a modular framework was developed for future engineering tools. The core of the framework consists of a module management including interfaces and a YANG-based datastore, maintaining a system-wide data model. Modules are provided for network configuration implementing CUC-functions, monitoring and container-based application and deployment.

关键词： Time-Sensitive Networking distributed real-time applications engineering framework

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：