TY - CHAP A1 - Wittig, M. A1 - Rütters, René A1 - Bragard, Michael ED - Reiff-Stephan, Jörg ED - Jäkel, Jens ED - Schwarz, André T1 - Application of RL in control systems using the example of a rotatory inverted pendulum T2 - Tagungsband AALE 2024 : Fit für die Zukunft: praktische Lösungen für die industrielle Automation N2 - In this paper, the use of reinforcement learning (RL) in control systems is investigated using a rotatory inverted pendulum as an example. The control behavior of an RL controller is compared to that of traditional LQR and MPC controllers. This is done by evaluating their behavior under optimal conditions, their disturbance behavior, their robustness and their development process. All the investigated controllers are developed using MATLAB and the Simulink simulation environment and later deployed to a real pendulum model powered by a Raspberry Pi. The RL algorithm used is Proximal Policy Optimization (PPO). The LQR controller exhibits an easy development process, an average to good control behavior and average to good robustness. A linear MPC controller could show excellent results under optimal operating conditions. However, when subjected to disturbances or deviations from the equilibrium point, it showed poor performance and sometimes instable behavior. Employing a nonlinear MPC Controller in real time was not possible due to the high computational effort involved. The RL controller exhibits by far the most versatile and robust control behavior. When operated in the simulation environment, it achieved a high control accuracy. When employed in the real system, however, it only shows average accuracy and a significantly greater performance loss compared to the simulation than the traditional controllers. With MATLAB, it is not yet possible to directly post-train the RL controller on the Raspberry Pi, which is an obstacle to the practical application of RL in a prototyping or teaching setting. Nevertheless, RL in general proves to be a flexible and powerful control method, which is well suited for complex or nonlinear systems where traditional controllers struggle. KW - Rotatory Inverted Pendulum KW - MPC KW - LQR KW - PPO KW - Reinforcement Learning Y1 - 2024 SN - 978-3-910103-02-3 U6 - http://dx.doi.org/10.33968/2024.53 N1 - 20. AALE-Konferenz. Bielefeld, 06.03.-08.03.2024. (Tagungsband unter https://doi.org/10.33968/2024.29) SP - 241 EP - 248 PB - le-tex publishing services GmbH CY - Leipzig ER - TY - CHAP A1 - Altherr, Lena A1 - Döring, Bernd A1 - Frauenrath, Tobias A1 - Groß, Rolf A1 - Mohan, Nijanthan A1 - Oyen, Marc A1 - Schnittcher, Lukas A1 - Voß, Norbert ED - Reiff-Stephan, Jörg ED - Jäkel, Jens ED - Schwarz, André T1 - DiggiTwin: ein interdisziplinäres Projekt zur Nutzung digitaler Zwillinge auf dem Weg zu einem klimaneutralen Gebäudebestand T2 - Tagungsband AALE 2024 : Fit für die Zukunft: praktische Lösungen für die industrielle Automation N2 - Im Hinblick auf die Klimaziele der Bundesrepublik Deutschland konzentriert sich das Projekt Diggi Twin auf die nachhaltige Gebäudeoptimierung. Grundlage für eine ganzheitliche Gebäudeüberwachung und -optimierung bildet dabei die Digitalisierung und Automation im Sinne eines Smart Buildings. Das interdisziplinäre Projekt der FH Aachen hat das Ziel, ein bestehendes Hochschulgebäude und einen Neubau an klimaneutrale Standards anzupassen. Im Rahmen des Projekts werden bekannte Verfahren, wie das Building Information Modeling (BIM), so erweitert, dass ein digitaler Gebäudezwilling entsteht. Dieser kann zur Optimierung des Gebäudebetriebs herangezogen werden, sowie als Basis für eine Erweiterung des Bewertungssystems Nachhaltiges Bauen (BNB) dienen. Mithilfe von Sensortechnologie und künstlicher Intelligenz kann so ein präzises Monitoring wichtiger Gebäudedaten erfolgen, um ungenutzte Energieeinsparpotenziale zu erkennen und zu nutzen. Das Projekt erforscht und setzt methodische Erkenntnisse zu BIM und digitalen Gebäudezwillingen praxisnah um, indem es spezifische Fragen zur Energie- und Ressourceneffizienz von Gebäuden untersucht und konkrete Lösungen für die Gebäudeoptimierung entwickelt. KW - Anomalieerkennung KW - IoT KW - Überwachung & Optimierung KW - DiggiTwin KW - BIM KW - Smart Building KW - Digitalisierung Y1 - 2024 SN - 978-3-910103-02-3 U6 - http://dx.doi.org/10.33968/2024.67 N1 - 20. AALE-Konferenz. Bielefeld, 06.03.-08.03.2024 (Tagungsband unter https://doi.org/10.33968/2024.29) SP - 341 EP - 346 PB - le-tex publishing services GmbH CY - Leipzig ER - TY - CHAP A1 - Hofmann, Till A1 - Limpert, Nicolas A1 - Mataré, Viktor A1 - Ferrein, Alexander A1 - Lakemeyer, Gerhard T1 - Winning the RoboCup Logistics League with Fast Navigation, Precise Manipulation, and Robust Goal Reasoning T2 - RoboCup 2019: Robot World Cup XXIII. RoboCup Y1 - 2019 SN - 978-3-030-35699-6 U6 - http://dx.doi.org/10.1007/978-3-030-35699-6_41 N1 - Lecture Notes in Computer Science, vol 11531 SP - 504 EP - 516 PB - Springer CY - Cham ER - TY - CHAP A1 - Hoegen, Anne von A1 - Doncker, Rik W. De A1 - Rütters, René T1 - Teaching Digital Control of Operational Amplifier Processes with a LabVIEW Interface and Embedded Hardware T2 - The 23rd International Conference on Electrical Machines and Systems (ICEMS), Hamamatsu, Japan Y1 - 2020 U6 - http://dx.doi.org/10.23919/ICEMS50442.2020.9290928 SP - 1117 EP - 1122 ER - TY - CHAP A1 - Hoegen, Anne von A1 - Doncker, Rik W. De A1 - Bragard, Michael A1 - Hoegen, Svenja von T1 - Problem-Based Learning in Automation Engineering: Performing a Remote Laboratory Session Serving Various Educational Attainments T2 - 2021 IEEE Global Engineering Education Conference (EDUCON) Y1 - 2021 U6 - http://dx.doi.org/10.1109/EDUCON46332.2021.9453925 SP - 1605 EP - 1614 ER - TY - CHAP A1 - Hüning, Felix A1 - Stüttgen, Marcel T1 - Work in Progress: Interdisciplinary projects in times of COVID-19 crisis – challenges, risks and chances T2 - 2021 IEEE Global Engineering Education Conference (EDUCON) Y1 - 2021 U6 - http://dx.doi.org/10.1109/EDUCON46332.2021.9454006 SP - 1175 EP - 1179 ER - TY - CHAP A1 - Altherr, Lena A1 - Ederer, Thorsten A1 - Schänzle, Christian A1 - Lorenz, Ulf A1 - Pelz, Peter F. T1 - Algorithmic system design using scaling and affinity laws T2 - Operations Research Proceedings 2015 N2 - Energy-efficient components do not automatically lead to energy-efficient systems. Technical Operations Research (TOR) shifts the focus from the single component to the system as a whole and finds its optimal topology and operating strategy simultaneously. In previous works, we provided a preselected construction kit of suitable components for the algorithm. This approach may give rise to a combinatorial explosion if the preselection cannot be cut down to a reasonable number by human intuition. To reduce the number of discrete decisions, we integrate laws derived from similarity theory into the optimization model. Since the physical characteristics of a production series are similar, it can be described by affinity and scaling laws. Making use of these laws, our construction kit can be modeled more efficiently: Instead of a preselection of components, it now encompasses whole model ranges. This allows us to significantly increase the number of possible set-ups in our model. In this paper, we present how to embed this new formulation into a mixed-integer program and assess the run time via benchmarks. We present our approach on the example of a ventilation system design problem. KW - Optimal Topology KW - Piecewise Linearization KW - Ventilation System KW - Similarity Theory Y1 - 2017 SN - 978-3-319-42901-4 SN - 978-3-319-42902-1 U6 - http://dx.doi.org/10.1007/978-3-319-42902-1 N1 - International Conference of the German, Austrian and Swiss Operations Research Societies (GOR, ÖGOR, SVOR/ASRO), University of Vienna, Austria, September 1-4, 2015 SP - 605 EP - 611 PB - Springer CY - Cham ER - TY - CHAP A1 - Nikolovski, Gjorgji A1 - Reke, Michael A1 - Elsen, Ingo A1 - Schiffer, Stefan T1 - Machine learning based 3D object detection for navigation in unstructured environments T2 - 2021 IEEE Intelligent Vehicles Symposium Workshops (IV Workshops) N2 - 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. KW - 3D object detection KW - LiDAR KW - autonomous driving KW - Deep learning KW - Three-dimensional displays Y1 - 2021 SN - 978-1-6654-7921-9 U6 - http://dx.doi.org/10.1109/IVWorkshops54471.2021.9669218 N1 - 2021 IEEE Intelligent Vehicles Symposium Workshops (IV Workshops), 11-17 July 2021, Nagoya, Japan. SP - 236 EP - 242 PB - IEEE ER - TY - CHAP A1 - Altherr, Lena A1 - Ederer, Thorsten A1 - Vergé, Angela A1 - Pelz, Peter F. T1 - Algorithmische Struktursynthese eines hydrostatischen Getriebes T2 - Antriebssysteme 2015 : Elektrik, Mechanik, Fluidtechnik in der Anwendung Y1 - 2015 SN - 978-3-18-092268-3 N1 - Antriebssysteme 2015 - Elektrik, Mechanik, Fluidtechnik in der Anwendung. VDI/VDE-Fachtagung. 11.11.15-12.11.15, Aachen. Veröffentlicht in der Reihe VDI-Berichte, Bandnummer 2268. SP - 145 EP - 155 PB - VDI-Verlag CY - Düsseldorf ER - TY - CHAP A1 - Schänzle, Christian A1 - Altherr, Lena A1 - Ederer, Thorsten A1 - Lorenz, Ulf A1 - Pelz, Peter F. T1 - As good as it can be: Ventilation system design by a combined scaling and discrete optimization method T2 - Proceedings of FAN 2015 N2 - The understanding that optimized components do not automatically lead to energy-efficient systems sets the attention from the single component on the entire technical system. At TU Darmstadt, a new field of research named Technical Operations Research (TOR) has its origin. It combines mathematical and technical know-how for the optimal design of technical systems. We illustrate our optimization approach in a case study for the design of a ventilation system with the ambition to minimize the energy consumption for a temporal distribution of diverse load demands. By combining scaling laws with our optimization methods we find the optimal combination of fans and show the advantage of the use of multiple fans. Y1 - 2015 N1 - Proceedings of FAN 2015, Lyon (France), 15 – 17 April 2015 SP - 1 EP - 11 ER -