Refine
Year of publication
Document Type
- Conference Proceeding (230) (remove)
Language
- English (230) (remove)
Keywords
- Enterprise Architecture (5)
- Engineering optimization (2)
- MINLP (2)
- Machine Learning (2)
- Robotic Process Automation (2)
- Serious Game (2)
- Ventilation System (2)
- Water distribution system (2)
- autonomous driving (2)
- 3D object detection (1)
- Agile development (1)
- Agility (1)
- Android (1)
- Angle Sensor (1)
- Anomaly detection (1)
- Artificial Intelligence (1)
- Automation (1)
- Awareness (1)
- Building Automation (1)
- Business Engineering (1)
Institute
- Fachbereich Elektrotechnik und Informationstechnik (230) (remove)
Digital twins enable the modeling and simulation of real-world entities
(objects, processes or systems), resulting in improvements in the associated value
chains. The emerging field of quantum computing holds tremendous promise for
evolving this virtualization towards Quantum (Digital) Twins (QDT) and
ultimately Quantum Twins (QT). The quantum (digital) twin concept is not a
contradiction in terms - but instead describes a hybrid approach that can be
implemented using the technologies available today by combining classical
computing and digital twin concepts with quantum processing. This paper
presents the status quo of research and practice on quantum (digital) twins. It also
discuses their potential to create competitive advantage through real-time
simulation of highly complex, interconnected entities that helps companies better
address changes in their environment and differentiate their products and
services.
In this article we describe an Internet-of-Things sensing device with a wireless interface which is powered by the oftenoverlooked harvesting method of the Wiegand effect. The sensor can determine position, temperature or other resistively measurable quantities and can transmit the data via an ultra-low power ultra-wideband (UWB) data transmitter. With this approach we can energy-self-sufficiently acquire, process, and wirelessly transmit data in a pulsed operation. A proof-of-concept system was built up to prove the feasibility of the approach. The energy consumption of the system is analyzed and traced back in detail to the individual components, compared to the generated energy and processed to identify further optimization options. Based on the proof-of-concept, an application demonstrator was developed. Finally, we point out possible use cases.
One central challenge for self-driving cars is a proper path-planning. Once a trajectory has been found, the next challenge is to accurately and safely follow the precalculated path. The model-predictive controller (MPC) is a common approach for the lateral control of autonomous vehicles. The MPC uses a vehicle dynamics model to predict the future states of the vehicle for a given prediction horizon. However, in order to achieve real-time path control, the computational load is usually large, which leads to short prediction horizons. To deal with the computational load, the control algorithm can be parallelized on the graphics processing unit (GPU). In contrast to the widely used stochastic methods, in this paper we propose a deterministic approach based on grid search. Our approach focuses on systematically discovering the search area with different levels of granularity. To achieve this, we split the optimization algorithm into multiple iterations. The best sequence of each iteration is then used as an initial solution to the next iteration. The granularity increases, resulting in smooth and predictable steering angle sequences. We present a novel GPU-based algorithm and show its accuracy and realtime abilities with a number of real-world experiments.
Component failures within water supply systems can lead to significant performance losses. One way to address these losses is the explicit anticipation of failures within the design process. We consider a water supply system for high-rise buildings, where pump failures are the most likely failure scenarios. We explicitly consider these failures within an early design stage which leads to a more resilient system, i.e., a system which is able to operate under a predefined number of arbitrary pump failures. We use a mathematical optimization approach to compute such a resilient design. This is based on a multi-stage model for topology optimization, which can be described by a system of nonlinear inequalities and integrality constraints. Such a model has to be both computationally tractable and to represent the real-world system accurately. We therefore validate the algorithmic solutions using experiments on a scaled test rig for high-rise buildings. The test rig allows for an arbitrary connection of pumps to reproduce scaled versions of booster station designs for high-rise buildings. We experimentally verify the applicability of the presented optimization model and that the proposed resilience properties are also fulfilled in real systems.
The course Physics for Electrical Engineering is part of the curriculum of the
bachelor program Electrical Engineering at University of Applied Science Aachen.
Before covid-19 the course was conducted in a rather traditional way with all parts
(lecture, exercise and lab) face-to-face. This teaching approach changed
fundamentally within a week when the covid-19 limitations forced all courses to
distance learning. All parts of the course were transformed to pure distance learning
including synchronous and asynchronous parts for the lecture, live online-sessions
for the exercises and self-paced labs at home. Using these methods, the course was
able to impart the required knowledge and competencies. Taking the teacher’s
observations of the student’s learning behaviour and engagement, the formal and
informal feedback of the students and the results of the exams into account, the new
methods are evaluated with respect to effectiveness, sustainability and suitability for
competence transfer. Based on this analysis strong and weak points of the concept
and countermeasures to solve the weak points were identified. The analysis further
leads to a sustainable teaching approach combining synchronous and asynchronous
parts with self-paced learning times that can be used in a very flexible manner for
different learning scenarios, pure online, hybrid (mixture of online and presence
times) and pure presence teaching.
Communication via serial bus systems, like CAN, plays an important role for all kinds of embedded electronic and mechatronic systems. To cope up with the requirements for functional safety of safety-critical applications, there is a need to enhance the safety features of the communication systems. One measure to achieve a more robust communication is to add redundant data transmission path to the applications. In general, the communication of real-time embedded systems like automotive applications is tethered, and the redundant data transmission lines are also tethered, increasing the size of the wiring harness and the weight of the system. A radio link is preferred as a redundant transmission line as it uses a complementary transmission medium compared to the wired solution and in addition reduces wiring harness size and weight. Standard wireless links like Wi-Fi or Bluetooth cannot meet the requirements for real-time capability with regard to bus communication. Using the new dual-mode radio enables a redundant transmission line meeting all requirements with regard to real-time capability, robustness and transparency for the data bus. In addition, it provides a complementary transmission medium with regard to commonly used tethered links. A CAN bus system is used to demonstrate the redundant data transfer via tethered and wireless CAN.
In this paper we investigate the use of deep neural networks for 3D object detection in uncommon, unstructured environments such as in an open-pit mine. While neural nets are frequently used for object detection in regular autonomous driving applications, more unusual driving scenarios aside street traffic pose additional challenges. For one, the collection of appropriate data sets to train the networks is an issue. For another, testing the performance of trained networks often requires tailored integration with the particular domain as well. While there exist different solutions for these problems in regular autonomous driving, there are only very few approaches that work for special domains just as well. We address both the challenges above in this work. First, we discuss two possible ways of acquiring data for training and evaluation. That is, we evaluate a semi-automated annotation of recorded LIDAR data and we examine synthetic data generation. Using these datasets we train and test different deep neural network for the task of object detection. Second, we propose a possible integration of a ROS2 detector module for an autonomous driving platform. Finally, we present the performance of three state-of-the-art deep neural networks in the domain of 3D object detection on a synthetic dataset and a smaller one containing a characteristic object from an open-pit mine.
The initial idea of Robotic Process Automation (RPA) is the automation of business processes through a simple emulation of user input and output by software robots. Hence, it can be assumed that no changes of the used software systems and existing Enterprise Architecture (EA) is
required. In this short, practical paper we discuss this assumption based on a real-life implementation project. We show that a successful RPA implementation might require architectural work during analysis, implementation, and migration. As practical paper we focus on exemplary lessons-learned and new questions related to RPA and EA.