275003 research outputs found
Sort by
Iteratively reweighted least squares revisited Beitrag zur robusten Parametersch ätzung
No full text2334The least squares estimator is optimal for normally distributed measurement errors, but it can break down under gross measurement errors. M-estimators can take fat-tailed error distribution into account, which makes them more robust to gross measurement errors. In this paper, a simpler description of the classical theory of robust M-estimators is presented and used to describe M-estimators for uniformly distributed outliers. In addition, a family of well known robust loss functions is presented in this notation and connections to a kernel lifting method are shown, which can be used as an alternative to the usual IRLS algorithm for calculating the M-estimators
Quasi-static tensile and fatigue behaviour of a metastable austenitic stainless Cr-Ni-Cu-N steel
No full textThe tensile as well as the fatigue behaviour of an experimental copper-alloyed metastable austenitic stainless steel has been investigated. Three batches were studied: (i) cast material as reference state and two states were recrystallised after (ii) rotary swaging and (iii) forward impact extrusion. Deformation-induced α’-martensite was formed during both rotary swaging and forward impact extrusion with volume fractions of 48 vol% and 4 vol%, respectively. The material was then heat treated to form a fine-grained, fully austenitic microstructure. These fine-grained material states exhibited a higher strength compared to the cast state. However, strain-hardening was highest in the cast state. The effect of the initial microstructure on the cyclic deformation behaviour and fatigue life was investigated using strain-controlled tests. During cyclic deformation all three states exhibited a martensitic phase transformation and concurrently a pronounced cyclic hardening with the formation of up to approximately 70 vol% α’-martensite. The fine-grained state produced by rotary swaging and reversion annealing showed a significantly higher fatigue life of about factor 4 compared to the coarse-grained cast state. Relative to forward impact extrusion, rotary swaging provided a higher fatigue life and fatigue strength.13
Action Space Design in Reinforcement Learning for Robot Motor Skills
No full text40214032Practitioners often rely on intuition to select action spaces for learning. The choice can substantially impact final performance even when choosing among configuration-space representations such as joint position, velocity, and torque commands. We examine action space selection considering a wheeled-legged robot, a quadruped robot, and a simulated suite of locomotion, manipulation, and control tasks. We analyze the mechanisms by which action space can improve performance and conclude that the action space can influence learning performance substantially in a task-dependent way. Moreover, we find that much of the practical impact of action space selection on learning dynamics can be explained by improved policy initialization and behavior between timesteps
A SiGe J-Band Gilbert Cell-Based Frequency Doubler and Power Amplifier Chain with 10 dBm Output Power
No full text10791082In this paper, we present the design and measurement results of a 220-294 GHz frequency doubler and amplifier chain for radar applications. The frequency doubler is implemented using a bootstrapped Gilbert cell, while the subsequent amplifier consists of three stages, designed as differential cascodes with tail current sources. This architecture ensures a full differential signal chain, providing high common-mode rejection. Using different breakout MMICs, we present measurement results of the entire chain and individual components. For the entire chain, a maximum output power of 10 dBm with a 3 dB bandwidth of 74 GHz was measured around the center frequency of 255 GHz. The circuit is designed in IHP's SG13G3 130 nm SiGe BiCMOS technology, which features heterojunction bipolar transistors with an ft and fmax of 470 GHz / 650 GHz, respectively
Depressive and anxiety symptoms in individuals with Long-COVID: Does social network matter? - Results of a German Long-COVID study
No full textBackground: Some patients previously infected with SARS-CoV-2 develop symptoms that are summarized under the term Long COVID (LC). There is limited evidence on how social characteristics influence depressive and anxiety symptoms in adults with LC.
Methods: Depressive symptoms were assessed using the Center for Epidemiological Studies Depression Scale (CES-D). Anxiety symptoms were measured using the Generalized Anxiety Disorder Scale (GAD-7). The living situation (alone vs. with someone) was assessed by a single question. The social network was assessed using the Lubben Social Network Scale (LSNS). Multivariate regression analysis was performed.
Results: In n = 410 participants with LC (mean age = 47.12, SD = 12.24) from Leipzig, Germany, we found that living with others was associated with fewer depressive symptoms compared to living alone (B = -2.18, p = .011). A larger social network was associated with fewer anxiety symptoms in adults with LC symptoms (IRR = 0.99, p = .008).
Conclusion: Social factors, such as social network and living situation, are associated with mental health factors like depressive and anxiety symptoms in adults with LC. Further research is required in order to elucidate the complex interplay between social factors and mental health in people with LC in the context of longitudinal studies. Future research directions are discussed.40
Identification of surface roughness parameters for the function-oriented description of EDMed surfaces
No full text131136The surface properties and especially the surface roughness are important parameters to determine the functionality of components. It influences relevant properties such as corrosion resistance, wear resistance or tribological behaviour. The surface roughness is commonly described by using the arithmetic average roughness Ra. The average roughness provides a general measure of the height of the texture across a surface and indicates the quality of the surface and the properties. Due to the integral approach, the arithmetic average roughness cannot consider different surface topographies caused by different crater geometries generated during EDM. Therefore, two non-identical surfaces can have the same average roughness Ra. In order to overcome this inadequate description on EDM-generated surfaces, this work aims to identify surface parameters that can be used to describe these surfaces in a functionally oriented manner. To do so, different surface roughness values will be investigated to identify the most relevant parameters to describe the EDM‘ed surfaces. Furthermore, drop shape analysis will be conducted to identify the most influencing surface parameters on the contact angle between a liquid and the surface, which quantifies the wettability of the components surface
Mobile robots in ports - Inspections, manipulation and other use cases
No full text168176Die Verbreitung von mobilen Robotern nimmt zu - autonome Robotersysteme unterstützen die Menschen unterschiedlich in vielen Bereichen. In Häfen als Drehscheiben des internationalen Handels existiert eine Vielfalt von Arbeitssaufgaben, die derzeit größtenteils manuell umgesetzt werden. Viele dieser Aufgaben sind durch den repetitiven Charakter, gefährliches Arbeitsumfeld und zeitgleich hohe Anforderungen an die Prozessstabilität und -kosten gekennzeichnet. Dieser Artikel fasst die Herausforderungen und Trends in Häfen bezogen auf die Automatisierung zusammen und identifiziert drei Haupteinsatzfelder für einen erfolgsversprechenden Einsatz von mobilen Robotern: Transport, Bestandsaufnahmen und Manipulation. Für jede dieser Haupteinsatzfelder werden anwendungsorientierte Beispiele vorgestellt.
AbstractThe use of mobile robots is increasing - autonomous robots support people differently in a wide range of areas. In ports, which are hubs of international trade, there is a variety of tasks that are currently performed manually. Many of these tasks are characterized by their repetitive nature, dangerous working environment, and high demands on process stability and cost efficiency. This article summarizes the challenges and trends in ports with regard to automation and identifies three main application areas for the promising use of mobile robots: transport, inspections, and manipulation. Application-oriented examples are presented for each of these main application areas.74
GreenPipe: Energy-Efficient Data-Processing Pipelines for Resource-Constrained Systems
No full textBillions of resource-constrained systems, such as embedded devices and cyber-physical systems, are in operation worldwide. These systems process input data (e.g., sensor data) into control signals for actuators or human-readable information, thereby providing valuable services and insights. Modern software methods, such as machine learning, have the potential to enhance the performance of these systems even further. However, machine learning is often associated with excessive energy demand, which urgently needs to be resolved. To address this issue, we present GreenPipe, an approach that creates energy-efficient data-processing pipelines tailored for embedded systems known for their low power demand. GreenPipe combines traditional AutoML techniques with energy models and thereby enables the selection of energy-efficient and accurate data-processing pipelines. We implemented GreenPipe on an ARM Cortex-M4 platform and evaluated its performance and energy efficiency. We demonstrate GreenPipe’s capabilities through a comprehensive evaluation, including a practical realworld application for predicting machinery-bearing faults. Green-Pipe demonstrates that it can reduce the energy footprint by up to 90% while maintaining high accuracy
Enhanced Flexibility in Direct Laser Interference Patterning through Industrial Robot Integration
No full text247254Direct Laser Interference Patterning (DLIP) is a well-established method for surface functionalization, enhancing properties like friction, wetting, and bacteria repellence. Traditionally, DLIP uses transla-tional stage or scanner systems to move either the laser or the component, offering high processing velocities and precision. However, these methods face limitations with heavy 3D parts, such as form-ing tools, reducing the range of components and applications. The study presents a new DLIP ap-proach using an industrial robot, providing more freedom for treating large components with complex geometries. The DLIP tool, including the laser source, is attached to a robot arm. The laser delivers nanosecond pulses with up to 120 W power and 40 mJ pulse energies, designed to be lightweight and vibration-tolerant. The DLIP optics generate an elongated rectangular beam with a long depth of focus. Initial experiments structured stainless steel plates with a pulse frequency up to 800 Hz and 15 W power, with structuring velocities from 2 to 20 mm/s. Different structuring strategies were applied to punches, and their tribological performance in forming applications was compared to non-structured punches. The produced topographies are characterized using a confocal microscope.19
Accuracy Study on Joint Source and Sensor Localization Using DOA and TDOA Measurements with Clock Biases
No full textThis paper addresses the problem of joint sensor and signal source localization using direction-of-arrival (DOA) and time-difference-of-arrival (TDOA) measurements from a sensor network, with an emphasis on accounting for sensor clock biases. The only requirements are that the position of one sensor is known in order to resolve ambiguities with regard to the position displacements and that the orientations of the sensors are known. Clock biases are incorporated into both the measurement model and the estimation process. The CramérRao Bound (CRB) is derived for the joint localization problem. We propose an estimation strategy based on an alternating optimization approach which decouples the position and bias estimation subproblems. Simulation results show that the root mean square error (RMSE) of the estimations attains the CRB and that the synchronization offsets can be recovered accurately