Skip to main content

Über dieses Buch

Der MHI e.V. ist ein Netzwerk leitender Universitätsprofessoren aus dem deutschsprachigen Raum, die sowohl grundlagenorientiert als auch anwendungsnah in der Montage, Handhabung und Industrierobotik erfolgreich forschend tätig sind. Die Gründung der Gesellschaft erfolgte im Frühjahr 2012. Der MHI e.V. hat derzeit 20 Mitglieder, die über ihre Institute und Lehrstühle zurzeit ca. 1.000 Wissenschaftler repräsentieren.
Die übergeordnete Zielsetzung des MHI e.V. ist die Förderung der Zusammenarbeit von deutschsprachigen Wissenschaftlerinnen und Wissenschaftlern untereinander, sowie mit der Industrie im Bereich Montage, Handhabung und Industrierobotik zur Beschleunigung der Forschung, Optimierung der Lehre und zur Verbesserung der internationalen Wettbewerbsfähigkeit der deutschen Industrie in diesem Bereich.
Das Kolloquium fokussiert auf einen akademischen Austausch auf hohem Niveau, um die gewonnenen Forschungsergebnisse zu verteilen, synergetische Effekte und Trends zu bestimmen, die Akteure persönlich zu verbinden und das Forschungsfeld sowie die MHI-Gemeinschaft zu stärken.



Optimization of a multi-component assembly workstation

This paper describes an approach to optimize workstations for manual assembly systems with high product variances by applying a set of methods in a structured manner. The main goal is to optimize workstations in terms of their efficiency and ergonomics. Therefore, this contribution provides an extensive analysis in the field of assembly efficiency and ergonomics. Based on the results, optimized concept versions are designed and evaluated by simulation. The favored concepts are presented to assembly workers prior to a cardboard engineering workshop. In the workshop, assembly workers build and optimize their own workplace with cardboard material. The workshop results lead to a final layout for implementation. The methodology has been implemented using the example of terminal strip assembly. A digital e-ink labeling system has been applied to ensure an efficient and ergonomic material arrangement. The use of a holistic methodology has resulted in a measurably more efficient and ergonomic workstation, which also uses digital technologies to support these positive effects.
Matthias Linsinger, Martin Sudhoff, Ivan Safonov, Christian Großmann, Andreas Schreiber, Bernd Kuhlenkoetter

Vision-based Generation of Precedence Graphs

The current developments in robotics aim towards a usage in households and small and medium-sized enterprises. In this case it is necessary to coordinate human and robot by describing the task in a common model. Precedence graphs are a possible representation of such a model. The generation of these is tedious especially for non-experts.We contribute a vision-based approach which generates precedence graphs based on a world representation generated by an registered eye-in-hand camera. To achieve this, we describe a world representation based on boundary representation models and a corresponding object recognition method. The results are used to generate the precedence graph by calculating the AND/OR graph as an intermediate step. We evaluate our approach based on several scene models and on a real world application.
Dorian Rohner, Myriel Fichtner, Dominik Henrich

Service-oriented Information Model for the Model-Driven Control of Dynamically Interconnected Assembly Systems

Classical, statically linked production systems are increasingly reaching their limits due to changing requirements, such as the production of customer-specific products up to lot size 1. Dynamically interconnected assembly systems offer a possibility for customized production, which can also react quickly to changing environments. An essential component of systems designed in this way is the matching between the requirements of a product and the capabilities of all existing and potential assembly resources. This paper therefore describes the product- and resource-side modeling that the components of the assembly system involved need as a basis for decision-making. In addition, it outlines the provision of information as well as the communication of the system components.
Dennis Grunert, Sven Jung, Thomas Leipold, Amon Göppert, Guido Hüttemann, Robert Schmitt

Creating and Using Digital Twins within Simulation Environments

Due to the increasing requirements in production environments, simulation is becoming more and more important in order to reduce time, costs and energy. It is not only the planning tools but also the production systems themselves (e.g. cyber-physical systems) which require simulations. In most cases, however, there is a huge effort needed to implement simulation possibilities. Moreover, the reuse of implemented parts and models is commonly limited. To react to these restrictions, we present a method to create digital twins in a system-neutral way. Our work delivers a corresponding architecture to use these twins in different simulation contexts as well as it provides a proof of concept.
Jannis Stecken, Matthias Bartelt, Bernd Kuhlenkötter

A Service-oriented Robotic Manufacturing System: Lessons Learned from Participating in the World Robot Challenge 2018

In order to manage fast changing market and customer requirements, production systems need to become more and more flexible. Therefore, the World Robot Challenge fosters the development of new agile manufacturing systems and evaluates them under tough conditions. In this paper, we introduce our service-oriented robotic system with which we participated in the challenge. The distributed software architecture covers four components: Robot Controller, Vision Module, Configuration Server and Process Controller. In order to enable a fast hardware setup, we used one collaborative robot and developed additional jigs and tools to solve the tasks. During development and testing under competitive conditions, we experienced the immediate benefits and drawbacks of our distributed architecture and its individual components. As a result, we can improve the partitioning of our services. The object detection based on a Convolutional Neural Network had to deal with false classifications due to changing lighting conditions. The combination of rough position estimation by using a global camera and a precise calibrated wrist camera for the visual servoing has turned out to be an essential element.
Ole Kröger, Jan Guhl, Oliver Heimann, Manuel Katanacho, Carsten Niebuhr, Martin Rudorfer, Tayyip Özkaya, Jonas Wassermann, Johannes Hügle, Tessa Pannen, Jörg Krüger

Distributed Software Architecture for Type-Specific AGV Routing

The current trend towards mass customization confronts established mass production systems with specific challenges, as these are designed for the production of large quantities with a small number of variants. Future production layouts will have to provide a high degree of flexibility and changeability in order to be able to respond to the demand for individualized mass products. This leads to a production network in which the work schedule of each product and thus the sequence of the individual production resources are flexibly arranged without losing the cost advantages of mass production. The current design forms, as a monolithic AGV system or the use of flexible autonomous units, show specific weaknesses that limit the use of these solutions in a flexible production network. The paper describes the research approach to establish an intralogistics service-based software architecture as an intralogistics execution system, which contains the elementary functions of a cross-platform use of mobile autonomous systems. One key application is the implementation of typespecific routes, which enables the cross-platform digitalization of the workspace to improve the interoperability and efficiency in route and path planning of different transport solution within one scenario. The implementation of the approach shows the advantages of a cross-platform service structure in general and the routing service in particular.
Michael Scholz, Martin Degel, Jörg Franke

Control Architecture for Industrial Robotics based on Container Virtualization

Current robot control is a rather static, monolithic application, where the application data has to be programmed, before the actual process starts. For the automation of more dynamical tasks, additional real-time control loops for process control and a more flexible robot control are needed. Therefore, this paper introduces a robot control architecture for real-time and separate asynchronous communication, which is able to build up local real-time control loops as well as including massively scalable cloud components. In a proof-of-concept, a multibody dynamic simulation of a robot and a controller are virtualized as separate components. The real-time suitability of the architecture is evaluated in comparison with non-virtualized and a monolithic variant of the same application.
Christoph Hinze, David A. Tomzik, Armin Lechler, Xun W. Xu, Alexander Verl

Evaluation Environment for Control Design of Soft Pneumatic Actuators

The topic of human-robot collaboration has become an important part of research in recent years. With suitable sensors and intelligent software, the workspace of man and robot is moving ever closer together. Another possibility is the use of soft material robots for the human-robot collaboration. However, the nonlinear material behavior complicate the control of the robots. In this paper, the low-level control of soft pneumatic actuators is researched. Therefore, a test bench for the low-level control is developed. The components of the test bench like the valve, the tubes and the actuator are modeled. Unknown parameters of the model are identified with the particle swarm optimization. Thus, a simulation model in addition to the test bench is built. With the simulation and the test bench, it is possible to investigate control strategies with focus on robustness due to leakage and parameter change. To illustrate this, a linear control strategy is implemented and tested.
Jan Christoph Krause, Serhat Ibrahim, Annika Raatz

Towards a Soft Shoulder Support System

The shoulder is burdened in many industrial contexts and is particularly very susceptible to injuries, which often result in a lack of productivity, unemployability, an overall depression, or a general presentism of the workforce. The use of support systems can prevent such injuries in general. The benefit of especially soft systems is a high wear comfort as well as its integrability in or wearability under regular and protective workwear. But the support level is lower compared to a hard system, because the human body itself must absorb strains instead of a sustaining frame. Besides, the development of such a system generating supportive torques at certain human joints by transmitting tensile forces along the body is challenging. This paper presents different developed approaches for a mechanism of a soft shoulder support system. The anteversion of the arm as a common shoulder movement is considered. Moreover, results of a study with one functional prototype are presented and discussed. It has been shown that this prototype provides an average muscle relief of about 10 % for static tasks, which is assessed as more than sufficient for work with light tools in overhead positions.
Niclas Hoffmann, Robert Weidner, Tim Schubert, Jens P. Wulfsberg

A Retrofit Approach for Predictive Maintenance

New technologies are driving the advancing digitalization of both consumer and industrial environments. Especially in maintenance, new possibilities to predict malfunctions and failures arise by real-time analytics of relevant data streams. In the context of this paper, the development and implementation of a predictive maintenance system based on low-cost embedded systems and sensor technology from the consumer sector will be presented. This includes the implementation of the device, its connection to the Internet of Things (IoT) as well as the application of machine learning. The system developed will be demonstrated at a robotic cell at the Institute of Production Systems (IPS).
René Wöstmann, André Barthelmey, Nikolai West, Jochen Deuse

Industry 4.0 beyond the Factory: An Application to Forestry

Industries like paper, packaging or furniture heavily rely on wood as basic resource, making the forestry sector an important part of the economy. However, the woods are not only an economic resource, but also a public good with significant cultural and ecological importance. This results in numerous boundary conditions and regulations, which aim at ensuring sustainable use. In combination with a high degree of division of labor and highly decentralized work processes within the sector, this results in highly complex value-added chains with a large number of stakeholders. Numerous innovative technological solutions exist, that aim at solving individual problems, yet, an overarching approach and infrastructure allowing for the realization of integrated value added networks is still missing.
Based on the manufacturing industries experience with Industry 4.0, we present a new concept Forestry 4.0. Its technical realization involves connecting wood resources, data sets, existing and new hard- and software components as well as stakeholders into a novel Internet of Things, Services, and People in forestry. Its nodes are Digital Twins which describe their corresponding real world artifacts, predict control their behavior, and integrate them into surrounding systems processes. In this way, Forestry 4.0 aims at integrating currently isolated artifacts and at revealing the potential of new technological solutions and business models. Finally, we show first steps of implementing this concept within a multidomain simulation system and present a strategy for the concepts dissemination into the forestry sector.
Jan Reitz, Michael Schluse, Jürgen Roßmann

Concept of an automated framework for sheet metal cold forming

This article describes a method to replace the current manual cold forming process of curved panels in ship building by a fully automated process control using a conventional press and crane setup. Basing on a simplified geometrical model and a self-learning artificial neural network the sheet metal forming as well as the derived forming strategy can be predicted. To position the workpiece that is suspended by four trolleys to the service point a flatness-based control algorithm has been developed. The algorithm enables sway-free movements. To monitor the production process successively, two 3D laser scanners are used. Reference marks applied on the workpiece surface are enabling the alignment calculation within the work space of the automated press as well as the plate contour at different bending lines. The successive laser scans recalibrate the neural network used for the local forming prediction and for forming process planning. The higher-level control assures the real-time capability of the control algorithms and the processing of all sensor data. Sensor multiplicity is used to supervise mechanism security and retains the acquired information in an all-embracing database to be considered in planning of future forming processes. The resulting database replaces the many years of experience of the workers.
Tobias Handreg, Pascal Froitzheim, Normen Fuchs, Wilko Flügge, Michael Stoltmann, Christoph Woernle

A Quick-Check to Evaluate Assembly Systems’ HRI Potential

Human-robot interaction (HRI) offers new opportunities for assembly processes which are executed manually. This is made possible by the development of lightweight robots which can interact directly with humans. To plan and implement HRI scenarios, suitable workstations and subprocesses need to be identified. The Quick-Check is a tool whose application allows an objective and reliable evaluation of HRI potential of existing, manual assembly workstations to be carried out with minimal effort in time and costs. This is achieved by rating the assembly subprocesses according to criteria which typically facilitate or complicate the automation of tasks. This article explains the motivation for as well as the relevance of the Quick-Check’s development. Furthermore, the state of the art in identification of automation and HRI potential in workplaces is presented. Additionally, the established criteria and the tool’s application are detailed. Finally, the Quick-Check’s validation by application in university and industrial environments and its integration into a planning system for the development of an HRI assembly system developed within the framework of a research project is described.
Ann-Kathrin Ermer, Tatjana Seckelmann, André Barthelmey, Kai Lemmerz, Paul Glogowski, Bernd Kuhlenkötter, Jochen Deuse

Safe Human Robot Collision Avoidance with an End-Effector integrated 360° LiDAR System

In future factories, new potentials rise with the possibility of direct interaction between humans and robots. Thereby the safety of human workers while collaborating with machines is of utmost importance. This contribution focusses on safety aspects arising by the end effector for handling applications. Presented is an approach to perceive the environment by an integrated LiDAR system to prevent possible collisions by alternating the trajectory from the endeffectors perspective. Several different sensor systems are discussed and evaluated before the chosen design is examined. Furthermore, two different methods to ensure a safe distance between human, workpiece and gripper are evaluated. Accordingly, a suitable control architecture for the guidance of the robot through the end effector is presented.
Markus Schäfer, Marcel Büsching, Sven Coutandin, Jürgen Fleischer

Methodology for digital planning and virtual validation of retroactively automated assembly systems

Disruptive technologies such as lightweight robots, autonomous mobile robots as well as powerful sensor and control technologies enable flexible and economical automation of production processes. Therefore, they are suitable for retroactive automation of existing systems. During retrofit, the conversion or commissioning time is a critical factor, which is why exact planning and virtual validation are necessary. Novel, digital tools offer extensive assistance for this task. This article describes a systematic and application-oriented methodology for continuous digital planning and validation of these automation solutions up to a turnkey robotic automation system. A validation of the proposed methodology in an electronics production scenario is described.
Maximilian Metzner, Fabienne Barth, Jochen Bönig, Jörg Franke

Feasibility of Mutualistic Networks for Different Types of Assembly Organization

Manual assembly is typical for complex products with large-scales and low repetition rates. The assembly control is usually done by supervisors and takes up a large part of their working time. The approach for decentralized assembly control is used in the assembly planning for relieving the supervisors and for making the assembly process more efficient. Primarily, this approach is used in flow-lines and is based on bionic mutualistic networks, which coordinate the assembly tasks by using an artificial stimulus. This article presents a transfer concept of the approach of decentralized assembly control for different alternative types of assembly organization. For transferring the approach, apart from flowline production, key requirements are figured out and the existing types of organization are systemized. The results of an enhancement of the approach suggest that especially organizational types with aperiodic worker movements in combination with stationary or aperiodic motion structure of assembly objects are suitable for the decentralized assembly control. A first testing already verifies the qualification of the approach for team assembly with stationary products and aperiodic worker movements.
Torsten Sievers, Katharina Krist, Daniil Volf, Kirsten Tracht

Potential of Energy Storage Systems for Industrial Robots

This paper presents a new approach to estimate the benefit of a energy storage for certain robots. This method can be used directly in the planning phase of production. First, a robot model is developed including the DC grid coupling of the individual drives. This model is validated by several measurements of the absorbed power, brake power and DC grid voltage in a real car body shop. In a next step, the model is used to estimate the potential of an energy storage system for robots in a specific production. The e stimation was successfully validated with and without energy storage. In the experimental evaluation an energy saving of 10 % was achieved.
Elias Knöchelmann, Jens Kotlarski, Timo Böhm, Svenja Tappe, Tobias Ortmaier

Robot-guided, feature-based 3D printing strategies for carbon fiber reinforced plastic

In the following paper, different feature-based 3D printing strategy approaches are presented for a robot-guided 3D printer that operates with carbon fiber reinforced plastic. First, the properties and the limitations of the printing material – currently under development – are shown. This is followed by the description of a robot-guided preliminary experimental setup, which can be applied to create individual parts or patches. The setup is shown through the mechanical components, the communication protocol is presented and the programming process is described. Subsequently, various feature-based strategies are explained in detail while their possible fields of application are highlighted.
Zsolt Kállai, Thorsten Schüppstuhl

Investigation of a force-guided assembly strategy for radial mounting of jet engine blades

Increasing air traffic and growing pressure on costs in aviation MRO, especially for jet engines, demand for an increase of process efficiency. While research concentrates on inspection and repair methods, the potential for automated reassembly is thereby mostly neglected. An approach for automated assembly of compressor and turbine blades is introduced in [1]. Part of this approach is a force-guided assembly, which is performed by an industrial robot. The aim of this force-guidance is to compensate positioning deviations, which can occur at different steps throughout the assembly process. Furthermore, sticking of blades during tangential movement, as a consequence of tilting, needs to be avoided. In this paper the proposed force-guided assembly strategy is investigated concerning its capacity of fulfilling the two claims stated above. Therefore, the hardware setup which is used to conduct the assembly and the process execution are described, as well as sources of positioning deviations are identified. Furthermore, the implementation of the force-control on the specific type of robot controller is described. Applying this force-control multiple experiments with defined positioning deviations are conducted. Furthermore, variations of additional assembly parameters are taken into account. The subsequent evaluation will allow a comparison of deviations, occurring during assembly process and the ability to compensate them. Potential for further optimization is stated in the conclusion.
Maik Dammann, Thorsten Schüppstuhl

Releasing Strategies for Gecko-Inspired Microgrippers

The handling of micro parts presents a challenge for the grasping mechanism. With a reduction of size, the surface shrinks by the power of two, while the volume shrinks by the third power. This results in an increasing impact of surface effects like adhesion forces compared to the gravitational force. While this effect complicates the handling of micro parts, it also allows handling principals, which are not suitable in the macro domain. A way of implementing these Effects for handling tasks is the adoption of biomimetic structures, as found on the surface of gecko or bug feet. In order to utilize this material for micro parts, a releasing strategy is needed. In this paper, strategies for releasing parts from gecko-like grippers in the macro domain are investigated and the basic principles are derived. These are investigated for their applicability for the handling of microspheres and thin foils. A grasping device is devised and the feasibility is shown in a proof of concept.
Finn Meiners, Alexander Bader, Jannik Beerepoot, Kirsten Tracht

Best-fit method for the calibration of 3D objects using a laser line sensor mounted on the flange of an articulated robot

Industrial robots are often equipped with external sensors that help perform the required production tasks. These sensors may be optical, tactile or force-torque sensors. Depending on the sensor, the measuring range can cover a one-dimensional, two-dimensional or threedimensional space. More specifically, laser line sensors are often used for inspection and calibration of 3D objects. Implementing an intuitive calibration method of such sensors on the robot flange is a challenging task. This is due to the fact that robotic applications and measurements are usually implemented with the presence of noise. As a result, best-fit methods or filters are implemented in order to realize an effective usage of the sensors. This paper deals with the concept of calibrating laser line sensors in order to measure cylindrical components using a bestfit method. Moreover, a modular control system for configuration and operation of the station with a variety of interfaces is discussed.
Rainer Müller, Matthias Vette-Steinkamp, Tobias Masiak, Ali Kanso

Optimization of a P/PI Cascade Motion Controller for a 3-DOF Delta Robot

An auto-tuning method for a Delta robot’s P/PI cascade motion controller using multi-objective optimization algorithm is proposed. The implemented control structure consists of two controllers: A feedforward controller based on a model of the inverse dynamics of the robot, and a cascade P/PI controller to compensate for unmodeled effects. The auto-tuning is achieved in the sense of optimizing the control parameters in three stages. In the first stage, the feedback control parameters are optimized after neglecting the feedforward control term. The goal is to minimize the position error in tracking an excitation trajectory, which is used as well to identify the dynamic model parameters in the second stage. After that, the feedforward compensation term is computed offline based on the desired trajectory. In the final stage, the P/PI parameters are optimized again after adding the feedforward controller. Experimental results on an industrial 3-dof Delta robot validates the efficiency of the proposed method.
Ahmed Zidan, Daniel Kaczor, Svenja Tappe, Tobias Ortmaier

Assignment, Sequencing and Location Planning in Line-less Mobile Assembly Systems

Industrial assembly systems, in particular for large scale products manufactured in lot size 1, require increasing levels of flexibility and (re-)configurability to shorten response times in volatile market environments thus maintaining competitiveness. Expanding on existing approaches for flexible production systems, we introduce the organizational and technological principle of Line-less Mobile Assembly Systems (LMAS) by combining the principles of dynamic job routes and mobile resources. Beside product-station assignments and schedules, operations planning for LMAS requires the determination of favorable assembly locations on the shop floor. Consequently, we introduce a linearized multistaged cost-based decision model for LMAS operation.
Armin F. Buckhorst, Guido Huettemann, Lea Grahn, Robert H. Schmitt

Development of a mobile robot cell for cutting up rotor blades

Electricity generation from renewable resources is still a global growth market. Especially wind turbines provide a significant part of the electricity generation. In addition to further technological developments and rationalization in production and assembly, the dismantling and recycling of old wind turbines is gaining in importance. Due to the high logistical costs involved in transporting entire rotor blades, various processes are currently being developed to dismantle rotor blades to be scrapped directly on site. Conventional concepts are mainly based on sawing processes which may pollute the environment and endanger the health of operators with dust and vapors. In addition these processes create bulky sub-segments.
The method presented in this work tries to avoid these disadvantages by using an alternative cutting method. It combines water jet cutting with the flexibility and accuracy of an industrial robot. In this paper, both the plant concept of a mobile, robot-based cutting cell and the method for automatic robot programming using real geometric data of the work piece are presented.
Tobias Haberecht, André Harmel, Martin-Christoph Wanner, Oliver Wurst

Improving Self-Localization Using CNN-based Monocular Landmark Detection and Distance Estimation in Virtual Testbeds

Mobile robots often require knowledge of their precise position. However, in many cases the integration of an outside-in tracking system is not feasible. In the exemplary case of autonomous vehicles, Global Navigation Satellite Systems (GNSS) are available but do not fulfill the precision requirements sufficiently strictly. Monocular cameras are another technology which is already built into many commercially available vehicles to enable additional safety and comfort features. Their image streams can be utilized to enhance localization precision using prior knowledge of landmarks along the roads.
In this paper, an integrated architecture for landmark detection, classification, position estimation and landmark-based localization is presented. The system is developed within a virtual testbed allowing for rapid development and evaluation of the employed components. In testing, an accuracy in the order of 10 cm is achieved for localization, constituting a significant improvement compared to systems using positional data from navigation satellites only.
Alexander Atanasyan, Jürgen Roßmann

A Fast Robot Playback Programming System Using Video Editing Concepts

Programming industrial robots is usually costly, tedious, and only feasible by experts. Hence, these systems do not play a big role in small and medium sized enterprises. In this paper, we present a novel approach for designing a robot programming system with the playback programming paradigm inspired by video editing concepts. First of all, we present the related work within this field of study. Afterwards, the programming system with its general structure, its user interface, and its input devices is described in detail. At the end, we discuss and interpret the results of the evaluation, so that these outcomes can be taken into account for future improvements of the programming system.
Michael Riedl, Dominik Henrich

Factorial Formulation of Dynamic Models for Robot Arms

Dynamic models are essential in many stages of the life-cycle of a robot. At the design stage these are used to dimension motors and drives and to evaluate dynamic performance. If the desired characteristics cannot be reached, current formulation and implementation of dynamic models cause a trial and error type approach to increase performance by changing properties of the robot in the design stage. Models only compute the forces and torques but do not offer any information about the factors influencing t heir v alues. I n t his p aper modifications to the formulation of the Iterative Newton-Euler method for dynamic modeling is proposed together with a reference class structure for implementation. Based on the mechanical structure and mass properties of the robot the dynamic model offers information about the factors and their proportions included in the reaction and motor forces or torques.
Akos Csiszar, Frank Weiß, Alexander Verl
Weitere Informationen