Search

The ongoing transformation of the manufacturing landscape introduces new business opportunities for enterprises but also brings new challenges with it. Especially small- and medium-sized companies (SMEs) require an increasing effort to stay competitive. Data produced on the shop-floor can be harnessed to conduct analyses useful to plant operators, e.g., for optimization of production capabilities or for increasing plant security. Therefore, we propose a privacy-preserving edge-computing architecture to facilitate a platform for utilizing such applications. Our approach is motivated by requirements from SMEs in Germany, e.g., protection of intellectual property, and employs suitable privacy models. We demonstrate the viability of the proposed framework by evaluation of uses cases for machine chatter optimization and anomaly detection within plants.

Galvanometer laser scanners are high-performance rotary motors for optical applications and are often utilized in production engineering. The dynamic properties of current scanner systems do not always fulfill the desired requirements. The controler-based solutions that have been presented in the past years primarily concentrate on changes in operation points or simple trajectories and neglect to address general trajectory tracking. This becomes a crucial issue in many production processes. In this article, a modification of adaptive inverse control applied to scanner systems to improve the dynamics and simultaneously address parameter and environmental changes is presented. An online identification of the system via an LMS-algorithm delivers an adaptive model of the PID-stabilized galvanometer drives. An inversion based feed-forward controller is designed, to improve dynamics and suppress eigenmovements, e.g. overshooting. The performance is verified by an experimental setup in which the error is reduced by almost 90 % for the chosen example, compared to a state of the art scanner system.

Remote laser beam welding (RLW) and remote laser beam cutting (RLC) are technologies that are well suited for automated applications in industrial facilities due to their contactless mode of action. The main advantages of both processes are the high operation speed, good process quality and a tailored contour of the welding seams and cutting kerfs. For an optimal use of these processes the laser beam has to be guided accurately over the work piece. A common technical realization of such processes is the combination of a laser beam deflection unit, a laser scanner and an industrial robot. This setup represents a programming challenge, as the six degrees of freedom of the robot plus the two degrees of freedom of the laser scanner lead to a kinematic redundant system. However, the advantages of this combination, the large robot workspace and the fast and precise motion of the laser scanner make it a very attractive technology for improvement of welding and cutting processes. An automated programming system for the described setup is presented in this paper.

Innovation processes of technical products underlie dynamic company internal and external influencing factors, which require a high degree of flexibility. Part of this flexibility is an adaptive behavior towards technical changes, in general within development as well as within production. These cycles of adaptations and changes necessitate an active coordination of information flows between the relevant domains. Accordingly, information about changes, e.g. the reasons as well as the consequences of changes, is transmitted at various interfaces within the innovation process. This provides the basis for deciding about the implementation of changes and adaptations across the different domains. Research has addressed the support of information exchange within innovation processes and specifically between the domains of product development and production. Nonetheless, the dynamics of internal and external influencing factors as well as of the related information exchange are not reflected comprehensively in approaches to support information exchange between these domains. Therefore, this paper addresses shortcomings of existing approaches regarding dynamics of information exchange through a literature based elaboration of interfaces and information flows between strategic product planning, product development and production planning. In this context, the production planning is represented by manufacturing technology, production resource and production structure planning. Accordingly, this paper aims at supporting sound decision making on the implementation of cyclic changes and adaptations in development and production by providing a classification of communication interfaces and information flows across domains.

Increasing requirements in regard to reactivity, adaptability and traceability in production and by extension in the supply chain can be observed among products, processes and clients all over the product lifecycle. In order to cope with these challenges, new methods for an accurate and reliable production planning and control are necessary. The integration of modern sensor technologies presents a promising approach to enhance information management on the shop floor. Hence, current states of both resources and individual products can be captured. The paper presents the framework of a multi-agent-supported, product-based production control, in which the individual product acts as an additional element of control in the production system. Thus, a significant contribution to adaptive and closed-loop production systems can be achieved that are able to transfer current information to the respective planning and control levels.

Nowadays turbulent environment exposes global producing companies to many risks caused by uncertain parameters. Therefore, a correct evaluation of investments in factories including uncertainties becomes increasingly important. Existing monetary evaluation methods as the calculation of the Net Present Value (NPV) integrate only static parameters like wages, material costs or overhead costs. Most evaluating methods do not consider the development of those parameters, e.g. future developments of wages. In a correct factory-evaluation method, a multitude of influence parameters including uncertain parameters should be integrated. These uncertainties can be both quantitative and qualitative. In this paper, factors like the development of wage cost, quantities and currency exchange rates are defined as quantitative uncertainties. So-called weak factors like cultural differences, staff motivation or quality awareness are specified as qualitative uncertainties. Qualitative uncertainties are not yet integrated into monetary factory-evaluation methods. Moreover, the dependencies between uncertainties — e.g. the impact of staff motivation to their productivity — have not been considered. This paper therefore presents an approach on how to integrate both qualitative and quantitative uncertainties and their dependencies into a monetary evaluation method for investments in factories through the use of a Fuzzy-Evaluation-Net. This Evaluation-Net is based on Fuzzy-Logic and is similar to the structure of Artificial-Neural-Networks (ANNs).

Telemanipulative microassembly is a promising research area which has attracted a great deal of attention in recent years. In these systems, precise movements and precise positioning are necessary tasks when low batches or prototypes have to be manufactured. Regarding these tasks, different teleoperator velocities might be useful. An experiment with human participants was conducted in order to find the appropriate joystick velocity and whether this adjustment depends on the task type. The participants had to track different paths and position the end effector of the microassembly system precisely. As dependent variables the tracking and positioning accuracy and the task-completion times were recorded while varying the joystick sensitivity. The results showed that tracking accuracy was best in sagittal direction and a joystick velocity of 0.090 cm/sec comprised the best proportion between speed and accuracy. Considering the positioning task, joystick sensitivities between 0.090 and 0.125 cm/sec yielded the best human performance for accuracy and task-completion times.

A method for intuitive and efficient programming of industrial robots based on Augmented Reality (AR) is presented, in which tool trajectories and target coordinates are interactively visualized and manipulated in the robot's environment by means of laser projection. The virtual information relevant for programming, such as trajectories and target coordinates, is projected into the robot's environment and can be manipulated interactively. For an intuitive and efficient user input to the system, spatial interaction techniques have been developed, which enable the user to virtually draw the desired motion paths for processing a work piece surface, directly onto the respective object. The discussed method has been implemented in an integrated AR-user interface and has been initially evaluated in an experimental programming scenario. The obtained results indicate that it enables significantly faster and easier programming of processing tasks compared to currently available shop-floor programming methods.

The remote nature of telepresence scenarios can be seen as a strongpoint and also as a weakness. Although it enables the remote control of robots in dangerous or inaccessible environments, it necessarily involves some kind of communication mechanism for the transmission of control signals. This communication mechanism necessarily involves adverse network effects such as delay. Three mechanisms aimed at improving the effects of network delay are presented in this paper: (1) Motion prediction to partially compensate for network delays, (2) Force prediction to learn a local force model, thereby reducing dependency on delayed force signals, and (3) Haptic data compression to reduce the required bandwidth of high frequency data. The utilised motion prediction scheme was shown to improve operator performance, but had no influence on operator immersion. The force prediction decreased the deviation between the delayed and the expected forces, thereby stabilising the control loop. The developed haptic data compression scheme reduced the number of packets sent across the network by 86%