Virtual commissioning (VC) is an established method for the efficient qualification of machine tool control software. Despite the potential to reduce the time-to-market and increase product quality, the creation of simulation models for VC still remains an elaborate task. The reason for this mainly results from the lack of data interoperability between engineering and simulation tools. To address this problem, we introduce a process for automatic behavior simulation model generation that covers the inter- and intra-company exchange of data in a heterogeneous toolchain. The workflow combines information of supplier components with ECAD engineering data represented through open standards like the Asset Administration Shell (AAS) and the Functional Mockup Interface (FMI). The desired level of integration between different data models is reached through a knowledge graph leveraging the Resource Description Framework (RDF) serialization of the AAS. In addition, a domain ontology for VC is introduced that semantically enriches the AAS data for simulation model generation. The presented concept is successfully validated with engineering data of a machine tool for sheet metal processing. It is shown, that a simulation model can be generated from a circuit diagram with minimum user interaction.

The position and path accuracy of industrial robots is strongly affected by temperature. In particular, temperature-dependent friction and altered hysteresis and transmission errors in the gearboxes are the main causes of these deviations. Since standard industrial robots lack temperature sensors, it is proposed to estimate the gearbox lubricant temperature using observer structures. It is planned to apply the proposed lubricant temperature observer to provide the input for temperature-dependent error compensation for gearboxes in industrial robots. First, a two-state heat transfer model describing the heat transfer from the lubricant through the gearbox housing to the environment is described and identified. It is integrated into a Luenberger observer and a Kalman filter with a correction of the lubricant temperature estimation based on measurements from a housing temperature sensor. Temperature measurements were performed in an experimental setup with a cycloidal gear filled with grease, which is common for industrial robots. Due to its changing viscosity and changing thermal conductivity at different shear rates, it shows a less uniform thermal behavior than gear-oil. For the validation measurements, the developed observers showed a mean absolute error of 1.3 K.

Large machine tools with long feed axes commonly utilize electrically preloaded rack-and-pinion drives, where two redundant drives on one axis are preloaded against each other by a dedicated control loop. Manufacturing- and assembly-related tolerances and elastic deformations lead to transmission errors in the drivetrains, that limit the achievable accuracy. Due to the limited bandwidth of the position control, these cannot be fully suppressed despite the use of linear measuring systems. In this paper, the superposition of the individual errors of the two drivetrains of electrically preloaded systems and the transfer function to the resulting path errors are theoretically derived in a first step. In addition, the negative effect of transmission errors on the preload control is investigated. To address the issue, an error compensation for the position and preload control loops is implemented. The basis for this are models of the state-dependent transmission errors utilizing a two-stage machine learning approach. The achievable reduction of path errors is validated in experimental investigations on a test bench and averages 57.8% for the measurements conducted.

Human–robot collaboration (HRC) offers promising potential for more flexible and sustainable production practices in architecture and construction. This requires HRC setups to scale up from light-payload collaborative robots to conform with the scale of building construction while considering the safety and teamwork culture for workers. This research proposes a system for large-scale multi-user HRC using head-mounted augmented reality (AR) devices. To achieve this, we contribute three methods that work in conjunction: (1) an AR system that enables multiple users to share tasks and work together with robots; (2) a dynamic human task allocation engine that reacts to the changing production teams and task types; and (3) a safety zone generation and allocation method to configure human collaboration in shared space with large-scale robots. The system is evaluated using a case study of prefabricated timber cassettes combining discrete event simulations , a user study and a fabrication process demonstrator with an industry partner.

Large industrial robots are inexpensive compared to their workspace. Due to flexibility, the positioning accuracy is compromised. This paper presents a simple yet effective static calibration method for robots using a laser tracker that can be applied to processes with no change in external load and without high processing forces. The calibration method is based on a novel elastokinematic model that captures the nonlinear gearbox and link compliance and also accounts for the effects of pose dependent joint friction and nonlinearities due to the gravity compensation mechanism. Furthermore, a feedforward concept is presented that separates kinematics from compliance to reduce computation time. The calibration method was implemented on industrial hardware and experimentally validated on a KUKA KR210–2. In comparison to state of the art compliance modeling, accuracy is improved by 23.1 %. Finally, the feedforward model reduces the mean positioning error in the entire workspace by 54.9% compared to the purely kinematically calibrated model, emphasizing the need for compliance compensation.

Recent advances in AI research are boosting the relevance of realtime neural networks in automation technology. To facilitate the deployment of trained neural networks for realtime applications, where hard limits on upper cycle time are required, this paper contributed a two-fold approach: A Python library for parsing Keras sequential models to structured text using a single function-call, which can be run directly on industrial PLCs and a structured text library to encapsulate reoccurring code and to provide a readable, tested interface for the generated code. The resulting models are experimentally validated regarding their inference time and precision of the outputs.

A key feature of cloud computing is elasticity, which enables resources to be allocated and de-allocated according to dynamic application demands. The evolution of Industrial Control Systems (ICS) is moving towards a distributed converged infrastructure that incorporates computing and communication resources as well as software, replacing dedicated hardware running control applications. While modern ICS and cloud environments share similarities, the corresponding applications differ significantly in their execution models and timing requirements. To enhance the flexibility of ICS, control applications must adapt to changing requirements throughout operation, resulting in varying workloads. The potential for dynamic scaling in ICS is considerable, but is not yet being fully exploited. The objective of this paper is to examine the extent to which state-of-the-art cloud autoscaling concepts are applicable to the scaling of real-time workloads and to propose enhancements to these concepts in order to enable the implementation of dynamic auto scaling of real-time workloads. In conclusion, the fundamental concept of cloud autoscaling can be applied, but must be extended, in particular to ensure deterministic resource provisioning and to improve task isolation in order to guarantee predictable task timing behaviour.

The value added by virtual commissioning of in-dustrial production systems is evident through significant cost savings achieved via early error detection. Detecting errors early in the development process enhances project success. However, the development of simulation models is associated with manual effort, particularly during the early phases of the project when it comes to creating the first models. To reduce costs and increase automation in the redesign of simulation models, a novel concept is presented that automatically transforms unstructured customer requirements into initial model placeholders within a simulation environment. The approach employs a large language model to identify necessary processes and machines, estimating their dimensions and positioning them relative to each other. The output is then converted to a simulation model format and imported into a simulation environment for further refinement. To verify the solution, a synthetic expert, trained using a large language model, learns the probability correlations between plant types, machines, and processes, thereby enabling a more accurate assessment of the resulting simulation model.

In today's global market, manufacturers must be highly adaptable, producing profitably smaller batches. Disin-tegrating hardware and software and embracing modern IT concepts like containerization and Service-oriented Architecture enable this flexibility. While offering rapid reconfiguration, these approaches add complexity. Orchestration tools aid in analyzing response times, workload mapping, and parameter selection for containerized real-time applications. However, current tools struggle with managing distributed, heterogeneous, real-time apps. This paper proposes an orchestration solution based on Compositional Performance Analysis and Genetic Algorithm to optimize task allocation across diverse computing resources. The approach enables the orchestration of heterogeneously scheduled, container-based. distributed real-time systems.

Distributed real-time applications perform deterministic tasks across multiple connected compute platforms, referred to as endpoints. Deterministic behavior of the communication and the compute are prerequisites, but coordinating and managing both in a holistic way is a significant challenge. Converged real-time networks based on Time-Sensitive Networking (TSN) are considered the key enabler across industries for the required connectivity. To date, the potential of TSN cannot be exploited due to many sub-standards still being under development, gaps in the management and configuration functions, lack of best practices and unresolved integration with higher application layers and domain specific engineering. Previous work, focusing on a modular open source framework for the engineering and management of distributed real-time applications, identified a unified approach for TSN-based endpoints as a key gap. This paper presents a novel endpoint architecture to operate, engineer and manage distributed real-time applications by seamlessly integrating real-time networking and compute. The architecture provides interfaces to a central engineering, allows to integrate centralized as well as distributed network management and offers a unified data management. An abstraction layer for applications allows the efficient use and complements existing open source solutions.

The increasing digitization of the manufacturing industry offers a wide range of opportunities for the effective use of production data, which can lead to significant time and cost savings. One promising approach is the use of synchronous simulations during the actual operation of cyber-physical production systems (CPPS) to provide valuable insights into possible performance improvements and at the same time reduce the risk of costly downtime. This paper presents a novel concept that enables structured storage of CPPS production data as part of a digital twin that serves as a virtual replica of the physical production environment. In particular, this approach is based on the Asset Administration Shell (AAS), which provides a comprehensive framework for the management of production data and allows for a seamless integration with simulation models. Storing production data in a structured and standardized format can make data queries efficient, facilitate data processing, and provide a solid foundation for simulation model development, calibration, and validation. These simulation models can then be used to identify potential process improvements and test different scenarios, minimizing the risk of disruption to actual production.

Static configurations and slow adaption to changing requirements characterize today’s production systems. In order to cope with variable external influences, these systems need to become more flexible. One building block towards meeting this requirement from the software perspective is employing quality-preserving processes to develop and update automation applications. While this condition applies to classical automation applications, it is increased if Software-defined Manufacturing (SDM) is applied. Containers can be used to deploy software through orchestration. Orchestration covers initial deployment and the roll-out of new software versions. This makes it in theory possible to deploy automation applications, such as computerized numerical control (CNC) software. However, available tools from information technology cannot be used to orchestrate and update automation applications, as real-time requirements must be considered. This work presents an approach that supports orchestration, specifically the deployment and update process, of real-time containers safeguarded by real-time capable simulation coupled with the control system during the operation phase. It shows how to achieve the right timing for real-time software updates and how to prevent causing damage by observation and safeguarding through information obtained from the real-time simulation. Using real-time simulation in the operating phase, container-based control and real-time orchestration become safer and more applicable in industrial manufacturing systems.

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.

Industrial robots are now used in a wide range of manufacturing processes. The increasing demands on quality and the complexity of the processes cannot always be met by the range of functions that are available in numerical control systems. In the coreless winding of fiber-polymer composites, it is important to keep the roving tension as constant as possible while maintaining a high processing speed. Currently, the winding paths are programmed via a number of control points. This results in a trade-off between lower process properties and higher processing speed with fewer control points. Conversely, finer interpolation of the path can improve process characteristics at lower processing speeds. With interpolation methods that are adapted to the process, new possibilities open up to overcome these trade-offs. This paper presents a spline path planning algorithm based on clothoids. This algorithm considers both the process requirements and the dynamic characteristics of the manufacturing unit. Furthermore, the algorithm is validated on a test contour and compared with established commercial methods.

Automating complex tasks using robotic systems requires skills for planning, control and execution. This paper proposes a complete robotic system for maintenance automation, which can automate disassembly and assembly operations under environmental uncertainties (e.g. deviations between prior plan information). The cognition of the robotic system is based on a planning approach (using CAD and RGBD data) and includes a method to interpret a symbolic plan and transform it to a set of executable robot instructions. The complete system is experimentally evaluated using real-world applications. This work shows the first step to transfer these theoretical results into a practical robotic solution. Note to Practitioners—This paper is inspired by the actual deficit to automate maintenance tasks in manufacturing environments through robotic systems. The focus of this work is on providing a practical robotic solution which can automate maintenance tasks completely autonomously. For this, the different skills from manipulation planning to the control and execution are explained. One main aspect of this work is to show how to create elementary robot motion tasks from symbolic plans. The second important part is to explain the practical capabilities applied on real-world applications.

In the commissioning of industrial plants, it has been shown that the utilization of virtual commissioning reduces the overall commissioning time and at the same time increases the software quality of the plant control system. This is achieved by starting the development of the control software earlier in the plant development process, allowing more time for its development. Virtual commissioning is performed in advance of the commissioning of the physical plant using a simulation model of the plant. However, the transition from the simulated to the physical plant proves to be problematic if components or modules of the plant are not physically available. In these cases, the hybrid commissioning method, a hardware-in-the-loop simulation setup that includes both physical and simulated components, is desirable. This article first briefly discusses the disadvantages of virtual commissioning and the concept of hybrid commissioning of industrial plants. Then, aspects of an implementation of hybrid commissioning for PROFINET fieldbuses are discussed.

This article investigates the use of impuls transfer during mechanical impacts for the acceleration of feed axes. For this purpose, additional masses are attached to linear axes and moved by an actuator. Very high accelerations can be obtained by impacts with a defined velocity on contact surfaces. Thus, almost abrupt velocity changes can be achieved on the main drives. To enable driving through corners with constant path velocity, the concept consists of a cross table with two axis equiped with an impact actuator each. Investigations are carried out by using a reference contour and the results are compared with existing strategies for adjusting the corners.

For the commissioning of production plants, the virtual commissioning (VCOM) method has proven to accelerate the overall commissioning time by increasing the software quality of the plant controller. VCOM is carried out in advance of the commissioning of the real plant, using a simulated plant. However, the hard transition from the simulated to the real plant proves to be problematic if components of the real plant are not available. In these cases the hybrid commissioning (HCOM) method, a hardware-in-the-loop (HIL) simulation setup including both real and simulated components, is desirable. In this article, the disadvantages of a hard transition of VCOM, the concept of the HCOM and it's benefits for commissioning production plants are explained and discussed. Also, an evaluation of an implementation is briefly presented.

The digitalization of production is a strategic goal in the automation industry, though visions and industrial reality remain far apart. A key issue is that today's segmented communication architecture based on fieldbus networks does not fulfill the future requirements. Converged networks connecting IT and OT systems while maintaining deterministic guarantees for real-time communication are required instead. Time Sensitive Networking (TSN) in the context of IEEE 802.1 Ethernet is a key enabling technology for converged networks. The transition from current communication architectures towards converged networks will not be a disruptive change, but rather an evolutionary process involving many steps. In this paper, an analysis of different dimensions of this transition process and their evolutionary steps is presented. Potential migration paths from current to future production architectures and resulting intermediate brownfield scenarios are identified and analyzed with a focus on their communication architectures.

The handling of deformable linear objects (DLOs), such as cables or hoses, with industrial robots is hardly industrially automated due to the underactuated material behavior. Increasing computing power and decreasing cost of processors over the last decades are now allowing the online computation of more complex handling processes. Within this paper, DLOs are approximated as a multibody chain with a finite number of degrees of freedom. Based on that model, a nonlinear, partially IO-linearizing controller is derived to control an underactuated end point on a trajectory when grasped by a robot arm at the other end. The controller is verified in simulation for high dynamics and shown to significantly reduce the tracking error. A practical evaluation finally shows the qualitative suitability of the multibody approximation and is used to validate the functionality of the controller in an open-loop experiment.

Control and localization of deformable linear objects (DLOs) require models to handle their deformation. This paper proposes an approach to automatically generate a model from available visual sensor information. Based on point cloud data obtained from a 3D stereo camera, the kinematics of a multibody model formulation are derived. The approach aims to balance the tradeoff between computational complexity and model accuracy. This is achieved with a geometric error criterion that reduces the introduced degrees of freedom (DOF) of the model to a necessary minimum, representing the continuous shape with as few bodies as possible. The approach is evaluated analytically and validated with an experimental scenario of DLO manipulation.

Open Platform Communications Unified Architecture (OPC UA) is a relevant technology in the field of machine-to-machine communication that defines data transport protocols and standardizes information modeling. Instead of defining a generalized world model, domain-specific models enable interoperability. These domain-specific models can be developed and released as Companion Specifications. More than 40 Companion Specifications are currently in development or already released. In this paper the current approaches to generate an OPC UA information model or an OPC UA Companion Specification are analyzed. A new modeling tooling based on the Eclipse Modeling Framework (EMF) is presented. This tooling demonstrates the potential of a model-driven approach for the creation of the information model and the structured part of the descriptive text document of a Companion Specification.

Until now, reconfigurable automation systems require a high engineering effort for the reconfiguration in order to begin production. Plug-and-play concepts are rare and usually only cover part of the disciplines involved in the mechatronic system. One of the challenges of reconfigurable systems considered in the research project DEVEKOS is the rearrangement of positioning axes into simple kinematic structures as it is required for the manufacturing process. In order to implement synchronous control of the multi-axes group, the calculation of the forward and inverse kinematics is necessary. This paper introduces a framework in which the kinematic description of the multi-axes group can be imported in a customized AutomationML-file. The framework then verifies the characteristics of the kinematics and considers common strategies for solving the inverse kinematics. The implementation presented recognizes four different types of basic kinematic structures and automatically calculates the inverse kinematic equations in those cases.

Future production systems are expected to support dynamic plug-and-produce scenarios across different component manufacturers. Therefore, the concept of I4.0 components was developed. An I4.0 component describes an asset with an asset administration shell providing a standardized representation of its capabilities. Real-time I4.0 components additionally demand for a deterministic communication with bounded low latency, where the network topology is typically defined by physical links and infrastructure. In order to avoid rewiring and to support dynamic plug-and-produce scenarios with real-time I4.0 components, we present so-called virtual topologies decoupled from the physical network infrastructure. Thereby, we extend the concept of real-time I4.0 components to configure the required behavior and introduce an additional submodel for the administration shell. Moreover, a central SDN controller is enhanced to automatically configure the virtual topologies during runtime. These topologies are based on the new real-time communication technology IEEE 802.1 Time-Sensitive Networking (TSN). Finally, we evaluate the latency of the prototypical implementation within real-time I4.0 components depending on the frame size and network bandwidth.

If existing communication systems are converted to OPC UA, an OPC UA information model is required for the existing data structures. In small systems, an information model can be created manually with reasonable effort. In systems with numerous data structures, however, the effort for manual creation is too demanding. In order to continue using the existing data structures in the code without having to handle OPC UA-specific data structures, a transformation between the existing data structures and the OPC UA information model is necessary. This article presents a possibility of generating an OPC UA information model by a description of the existing data structures, e.g. generated based on ROS messages. The description of data structures is evaluated at runtime and linked with the existing data structures. The different representations for the different access types (read, request/response) in the OPC UA information model are discussed. In addition, an OPC UA client can read the information model from the descriptions of the data structures and write the read data back to the original data structure. In this way, both the provision of the data via OPC UA and the reception are possible. The adaptation effort of the existing software is minimal, since the existing structures are automatically populated and can still be used as usual.

Within the context of Industry 4.0, future production systems are expected to support mass customization up to lot size one. In order to realize the individual production processes, a changeable and consistent integration of assets from different manufacturers is required. Therefore, the concept of I4.0 components was developed. Every I4.0 component provides a so-called asset administration shell, i.e. a standardized virtual representation of its capabilities. This paper focuses on real-time I4.0 components that require a deterministic communication with bounded low-latency in-between. In contrast to non-real-time I4.0 components, Ethernet (IEEE 802.3) is enhanced by the new real-time communication technology Time-Sensitive Networking (TSN, IEEE 802.1). The basic configuration of TSN connections within the asset administration shell was already addressed by previous work. In this work, we present the configuration of application layer protocols using the example of OPC UA Publish/Subscribe (PubSub). More precisely, we outline the underlying concept and define additional submodels for the asset administration shell of real-time I4.0 components. Finally, we discuss the advantages of the presented approach in terms of changeability.

To increase the changeability of future production systems, assets have to provide a standardized representation of their capabilities. Therefore, the concept of I4.0 components was developed, which describes assets with an asset administration shell providing a self-description. Real-time I4.0 components, like safety critical systems or motion control applications, require a deterministic communication with bounded low latency. Moreover, they demand for redundant connections to achieve fault tolerance with respect to the network infrastructure and the components. The integration of redundancy mechanisms in realtime I4.0 components requires additional implementation and configuration effort, which has not been covered in detail yet. In this work, we extend the concept of asset administration shells to configure end-to-end redundancy between real-time I4.0 components. Thereby, we apply the new real-time communication technology Time-Sensitive Networking (TSN) with focus on the standard IEEE 802.1CB. Finally, we evaluate the performance of the redundancy implementation in terms of the transmission offset between the primary frame and its duplicate.

The need for temporal and spatial coordination of two robot arms moving independently in a shared workspace frequently arises in industrial and service-oriented robotics alike. Today, this problem is often solved manually, leading to a negative impact on user experience as well as on execution performance. In this paper, we present an algorithm that is able to automatically coordinate independently planned motions of a dual-arm manipulator during execution. In addition, the algorithm is capable of refining the plan upon receiving new motion commands during the robot motion. We demonstrate the effectiveness and efficiency of the proposed approach on an ABB YuMi robot working on an industrial palletizing task.

The research and development of intelligent automation solutions is a ground-breaking point for the factory of the future. A promising and challenging mission is the use of autonomous robot systems to automate tasks in the field of maintenance. For this purpose, the robot system must be able to plan autonomously the different manipulation tasks and the corresponding paths. Basic requirements are the development of algorithms with a low computational complexity and the possibility to deal with environmental uncertainties. In this paper, an approach is presented, which is especially suited to solve the problem of maintenance automation. For this purpose, offline data from CAD is combined with online data from an RGBD vision system via a probabilistic filter, to compensate uncertainties from offline data. For planning the different tasks, a method is explained, which uses a symbolic description, founded on a novel sampling-based method to compute the disassembly space. For path planning, we use global state-of-the-art algorithms with a method that allows the adaption of the exploration stepsize in order to reduce the planning time. Every method is experimentally validated and discussed.

High accurate absolute robot positioning is a requirement, and still a challenge, in many applications, such as drilling in the aerospace industry. The accuracy is affected due to many sources of errors from robot model, tool calibration, sensor and product uncertainties. While model-based error compensation cannot reach the desired accuracy, sensor-based compensation appears as the practical solution to increase the robot positioning accuracy. A structured analysis of the error sources in robotic manufacturing processes can facilitate error identification and further compensation. This paper describes an error source breaking down approach for analyzing robotic manufacturing processes. Moreover, an external sensor-based compensation is proposed for error reduction and error identification. Comparison with a compliance model-based compensation is performed. The proposed approach is applied to a robotic drilling process for aircraft manufacturing, considered a general and real industrial application. Further validation through experimentation is performed. The validation revealed a clear improvement in robot positioning accuracy and the benefits of the proposed error source structure for analysis.

In many robotic processes such as milling and drilling, there are multiple solutions for robot poses, as the rotation around the tool axis remains as a degree of freedom (DoF) in positioning. Yet until now, this DoF causes additional efforts in CAM programming as it requires manual intervention. Instead, this DoF can be used to optimize the robot pose according to different criteria of the robot such as stiffness or avoidance of backlash effects. This paper presents different criteria for optimization of the robot pose in machining and describes the optimization of robot stiffness based on a novel method for its identification, which is in detail described on this paper. Furthermore, the potential of the automatic resolution of the DoF is outlined enabling staff without robot knowledge to define reasonable robot paths.

Road and lane detection play an important role in autonomous driving and commercial driver-assistance systems. Vision-based road detection is an essential step towards autonomous driving, yet a challenging task due to illumination and complexity of the visual scenery. Urban scenes may present additional challenges such as intersections, multi-lane scenarios, or clutter due to heavy traffic. This paper presents an integrative approach to ego-lane detection that aims to be as simple as possible to enable real-time computation while being able to adapt to a variety of urban and rural traffic scenarios. The approach at hand combines and extends a road segmentation method in an illumination-invariant color image, lane markings detection using a ridge operator, and road geometry estimation using RANdom SAmple Consensus (RANSAC). Employing the segmented road region as a prior for lane markings extraction significantly improves the execution time and success rate of the RANSAC algorithm, and makes the detection of weakly pronounced ridge structures computationally tractable, thus enabling ego-lane detection even in the absence of lane markings. Segmentation performance is shown to increase when moving from a color-based to a histogram correlation-based model. The power and robustness of this algorithm has been demonstrated in a car simulation system as well as in the challenging KITTI data base of real-world urban traffic scenarios.

In robotic machining applications, the precision of the robot is of great importance. In heavy machining process, the lower stiffness of industrial robots results in greater position errors than that of the CNC machine executing the same process. In this contribution, a new stiffness model with 36 degrees of freedom and nonlinear descriptions are presented together with a new identification method. Experimental results outline the potential of the model in machining application.

At the core of every object recognition system lies the development and integration of distinct feature descriptors to create object representations robust against varying perspectives or lightning conditions. Recent work has primarily focused on the development of distinct point features. While these features achieve impressive recognition results, point features fail to capture the shape and appearance of an object with less or even without texture. This paper proposes a novel method for the rapid and dense computation of 2D and 3D image cues from RGB-D data to target the recognition of objects without rich texture and a global histogram-based descriptor for the distinct description of object models.

In recent years, driver-assistance systems have emerged as one major possibility to increase comfort and safety in road traffic. Still, cost is one major hindrance to the widespread use of safety systems such as lane-change or blind spot warning. To facilitate the widespread adoption of such safety systems, thus increasing safety for all traffic participants, the use of cost-efficient components is of crucial importance. Within this work we investigate the use of cost-efficient, widely used ultrasonic sensors for the tracking of passing-by vehicles at high velocities. Therefore, a particle filter with some mixture tracking capabilities is implemented to fuse the signals from 6 us-sensors. The main focus of this work lies on the development of a more detailed sensor model that is used in this particle filter. Additionally, a strategy to take into account object-visibility w.r.t. the different sensors is outlined. The derived concept is evaluated experimentally in real road traffic. The applicability of the tracking result in context of lane-change-decision-aid and blind-spot-surveillance systems is analyzed.

This paper investigates existing methods for 3D point feature description with a special emphasis on their expressiveness of the local surface geometry. We choose three promising descriptors, namely Radius-Based Surface Descriptor (RSD), Principal Curvatures (PC) and Fast Point Feature Histograms (FPFH), and present an approach for each of them to show how they can be used to classify primitive local surfaces such as cylinders, edges or corners in point clouds. Furthermore these descriptor-classifier combinations have to hold an in-depth evaluation to show their discriminative power and robustness in real world scenarios. Our analysis incorporates detailed accuracy measurements on sparse and noisy point clouds representing typical indoor setups for mobile robot tasks and considers the resource consumption to assure real-time processing.

Non-holonomic, omnidirectional undercarriages that are composed of steered standard wheels seem to provide a solid compromise between versatility, flexibility and high robustness against various ground conditions. However, such undercarriages are characterized by the occurrence of a number of singular configurations. To avoid these singular configurations most control-approaches restrict the admissible configuration-space thus eventually reducing the mobility and flexibility of the undercarriage. Within this work a state-space representation that forms a locally singularity-free atlas of the admissible configuration-space is presented. Based on this state-space description a switching based controller is developed that incorporates the former singular regions into the used configuration space and thus allows to exploit the full flexibility of non-holonomic, omnidirectional undercarriages. The implemented controller is quantitatively and qualitatively evaluated and compared to one approach that avoids the singular regions and one that completely neglects the non-holonomic bindings.

A growing number of modern computer vision applications like object recognition, collision avoidance and scene understanding demand accurate and dense 3D representations of their environment. To improve existing procedures for 3D data acquisition this paper proposes a novel method for sensor combination on a stereo and a Time-of-Flight camera system. By calibrating the two sensor systems to each other, valid measurements from the 2.5D Time-of-Flight sensor are converted to disparity guesses within the stereo system. The disparity guesses from the Time-of-Flight data constrain the correspondence search results from the stereo matching algorithm. It is empirically shown, that the proposed method effectively enhance the results from stereo vision, especially in structureless areas where stereo correspondence search fails. The method is evaluated on the camera system of the service robot Care-O-bot® 3.

State-of-the-art local feature descriptors like SIFT or SURF require a significant amount of computational power which prevents their usage in applications with real time constraints. Despite recent efforts to simplify the calculation of feature descriptors, a faster computation comes often to the disadvantage of weakening the invariance to rotation or scale. Recently, Tola et al. introduced DAISY, a new local feature descriptor for wide-baseline matching across stereo image pairs. It is shown that DAISY outperforms SIFT in terms of matching accuracy while being computed significantly faster. This paper takes on the idea of DAISY by proposing a rotational invariant extension of the descriptor, called O-DAISY, and outlining its implementation on FPGA to achieve real time performance. The results are benchmarked against its original version and against the widely used descriptors BRIEF and SURF on a standardized image set.

The integration of industrial robot systems into the manufacturing environments of small and medium sized enterprises is a key requirement to guarantee competitiveness and productivity. Due to the still complex and time-consuming procedure of robot path definition, novel programming strategies are needed, converting the robotic system into a flexible coworker that actively supports its operator. In this paper, a learning-from-demonstration strategy based on Hidden Markov Models is presented, which permits the robot system to adapt to user- as well as process-specific features. To evaluate the suitability of this approach for small-lot production, the learning strategy has been implemented for an arc welding robot and has been evaluated on-site at a medium sized metal-working company.

For the constantly growing service robotic market there is a demand for new energy-efficient and economically priced actuation-concepts. This paper describes the QuadHelix-Drive, a novel rope actuator of high power density with a simple working principle. It highlights the technical challenges, which evolved while examining the DoHelix-MuscIe-Concept. A strategy to overcome these challenges and a prototypic mechanical realization of this new actuator concept are iIIustrated. The integration of the QuadHelix-Drive into the Fraunhofer IPA testing facility is described and at the end possible robotic application scenarios are outIined.

Currently, pseudo-omnidirectional, wheeled mobile robots with independently steered and driven wheels seem to provide a solid compromise between complexity, flexibility and robustness. Yet, such undercarriages are imposed to the risk of actuator fighting and suffer from singular regions within their configuration space. To address these problems we expand a previously developed potential field (PF) based approach by expanding it with a predictive horizon. The proposed method is based on a model predictive control (MPC) approach, incorporating a gradient descent optimization step via the Pontryagin minimum principle. To enforce adherence to the constraints during optimization, we modify the Lagrange-multipliers within the backpropagation of the costates. The proposed approach is evaluated simulatively w.r.t. the undercarriage of the Care-O-bot®3 mobile robot and is compared to the potential field based and a model predictive control approach.

We present a practical strategy for real-time path planning for articulated robot arms in changing environments by integrating PRM for Changing Environments with 3D sensor data. Our implementation on Care-O-Bot 3 identifies bottlenecks in the algorithm and introduces new methods that solve the overall task of detecting obstacles and planning a path around them in under 100 ms. A fast planner is necessary to enable the robot to react to quickly changing human environments. We have tested our implementation in real-world experiments where a human subject enters the manipulation area, is detected and safely avoided by the robot. This capability is critical for future applications in automation and service robotics where humans will work closely with robots to jointly perform tasks.

For mobile platforms with steerable standard wheels it is necessary to precisely coordinate rotation and steering angle of their wheels. An established approach to ensure this is to represent the current state of motion in form of the Instantaneous Centre of Motion (ICM) and to derive a trajectory within this space. However, while control in the ICM space does guarantee adherence to the system's nonholonomic constraints, it does not avoid the system's singular configurations. Within this work we address the problem of singularity avoidance within the ICM space. Singularities related to the mathematical representation of the ICM are reduced by a reformulation of this representation. Furthermore, a controller based on artificial potential fields avoiding singular configurations of the robot by representing them as obstacles in the derived ICM space is designed. The resulting controller is particularized and analyzed w.r.t. the Care-O-bot 3 demonstrator.

This paper analyzes the potential to apply mobile service robots in offshore oil and gas producing environments. The required hardware and software components and abilities of such a mobile offshore inspection and manipulation robot are presented in this paper. Possible applications of mobile service robots in an offshore environment range from simple visual inspection tasks to physical intervention with the process equipment, e.g. for sample taking, valve turning, cleaning up minor obstructions, and operating control panels. The first prototype of a mobile offshore inspection robot is equipped with a robotic arm which carries a camera for visual inspection as well as various application sensors such as a microphone, gas and fire sensors. It is able of both, remote and autonomous inspection of industrial process equipment. In automatic mode the robot autonomously executes pre-programmed inspection tasks. The results of all inspection tasks are saved to a database and can be reviewed by the responsible operator in the central control room at any time. The evaluation of the first autonomous mobile robot that has ever been operated in offshore environments has proven the applicability of mobile robotics to offshore environments. Different types of inspection tasks (visual and acoustic inspection, gas measuring) have been programmed to and executed by the robot successfully without ever jeopardizing the safety of the platform or the platform personnel. The application of mobile robotics in offshore environments can reduce the level of manual human intervention required to operate a production facility thereby increasing the efficiency of the workforce, improving safety and working conditions, and improving the production economics. The successful evaluation of the first realization of a mobile inspection and manipulation robot has thus leveled the ground for future mobile robot installations in offshore environments.

This paper introduces Care-O-bot®3, a highly integrated and compact service robot with manipulation, navigation and vision capabilities. In particular, Care-O-bot®3 combines the best of available technology including a 7 DOF light-weight arm, an omnidirectional platform and many high-end sensors along with a sustainable, end user oriented design concept enabling many interaction possibilities.