128 phd tööd leitud asukohas Schwieberdingen

Die Feldorientierte Regelung (Field-Oriented Control, FOC) ist seit vielen Jahren das Standardverfahren für die Regelung von Synchron- und Asynchronmaschinen. Es sind allerdings auch Alternativen bekannt, wie die modellprädiktive Regelung (Model Predictive Control, MPC), die Vorteile z.B. bei der systematischen Behandlung von Stellgrößenbegrenzungen und bei der Berücksichtigung von produktrelevanten Zielgrößen haben. Ihre Aufgabe wird es sein, modellprädiktive Regelungen für permanenterregte Synchronmaschinen zu untersuchen. Ziel ist der Einsatz in der leichten Elektromobilität sowie in dynamischen Kleinantrieben. Schwerpunkte des Promotionsvorhabens sind: Erforschung und Entwicklung nichtlinearer modellprädiktiver Regelungsansätze (z.B. Direct MPC und Indirect MPC) für permanenterregte Synchronmaschinen, mit Fokus auf produktrelevante Zielgrößen wie z.B. ein verbessertes akustisches Geräuschverhalten (NVH) sowie reduzierte Verluste durch optimale Modulation. Hardwarenahe...

Field-Oriented Control (FOC) is since many years the state-of-the-art control strategy for synchronous as well asynchronous machines. However, Model Predictive Control (MPC) is gaining attention as an advantageous alternative due to its superior characteristics, such as the systematic handling of input as well as states constraints.  Your tasks will be to investigate MPC methods for permanent magnet synchronous machines (PMSMs). The aim is to investigate its possible use in light electromobility as well as in high-dynamic small drives. The doctoral thesis will focus on: Research and development of nonlinear MPC schemes (e.g. Direct MPC and Indirect MPC) for PMSMs, with focus on product-relevant Key Performance Indices (KPIs), such as improving the Noise, Vibration, and Harshness (NVH) behavior, or reducing the losses via optimal modulation. Hardware implementation of the investigated MPC schemes on Field Programmable Gate Arrays (FPGA) and on microcontrollers. Investigation of...

Our department is responsible for circuit-board technology development and is part of the European public funded project EECONE. In this context, development activities have a strong focus to re-think and re-develop the supply chain to achieve a circular and less resource intensive production to fulfill the company strategy towards CO2 emission neutrality. We are highly interested in gaining deeper knowledge and expertise on the system level regarding technologies for fabrication and patterning of circuit boards. You will join a team of scientists and engineers with many years of experience in the development of products for automotive applications. Your objective is to develop circuit-board and assembly and interconnection technology to fulfill the reliability requirements of the automotive industry with a strong focus on sustainability. The academic supervision from university side is currently under discussion. Task within the PhD work will be process development of...

The research team at the Bosch Center for Artificial Intelligence (BCAI) provides the opportunity for candidates currently enrolled as PhD student s  in machine learning, computer vision, or a related field to perform an 'industry sabbatical' for 3-12 months. At BCAI we build the foundation to generate real-world impact through cutting-edge research focusing on safe, robust, data-efficient, and explainable AI. We design and implement AI for smart, connected, and autonomous technologies across Bosch business sectors. During the industry sabbatical you will get the chance to develop and implement novel algorithms, and evaluate them on multi-modal open and Bosch datasets. Together with BCAI researchers you will perform original research, theoretical investigations, and aim to publish at top ML conferences. The goal is to define the topic in such a way that it is closely related to your PhD topic. Possible research areas include: Foundation models for perception (e.g., training...

As part of a collaborative research project, Robert Bosch GmbH and the University of Ulm aim to propel the technological progress in the development of cooperative radar sensor networks. Our goal is to significantly enhance sensor performance through a distributed arrangement without increasing the complexity of individual sensors. To this end, we are working on innovative concepts for networking millimeter-wave sensors with digital modulation. During the course of your PhD, you will research novel concepts and demonstrate their performance capabilities. Your tasks will involve bringing in your own ideas and verifying these through both simulations and measurements. Ideally, you will succeed in proving the merits of your sensor network approach in real-world scenarios. Your work will contribute to expanding the current state of the art and shaping the future of autonomous and connected vehicles. We look forward to your creative ideas and your dedication in solving this exciting...

Main goal in this PhD thesis is to develop a virtual design method for injection-molded thermoplastic parts by coupling mechanistic aging models to mechanical constitutive equations, enabling the virtualization of aging tests at the component level. You will identify pertinent aging models based on kinetic and diffusion laws. Furthermore, you will describe the evolution of relevant polymer descriptors as a function of different aging parameters as for example the temperature or the concentration of the aging agent. You will supplement and validate your findings by targeted experiments. Finally, you will link the mechanistic aging models to the mechanical constitutive equations and validate your approach on component level by simulation and actual aging tests. Qualifications Education : excellent degree in Mechanical/Aeronautical/Civil Engineering, Material Science, Physics, Applied Mathematics, or similar Experience and Knowledge : in depth knowledge of continuum...

The goal of the thesis is to develop methods for data-centric radar AI development, including relevant data selection, understanding of structures in data, recognition of invariances for easier generalization, and achieving robustness. The ongoing automation of the driving functions in cars needs accurate environment sensing and perception.  Radar is a key surround sensing technology for automated driving and AI based methods play a key role for radar perception. Ideally, the developed approaches are implemented prototypically and applied to a cutting edge radar AI use case in the field of autonomous vehicles.  Qualifications Education: excelent Master's degree in Data Science, Electrical Engineering, Physics or Applied Informatics                    Personality:  good communication skills Experience and Knowledge: good knowledge in state-of-the-art AI methods, digital signal processing, experience in programming in python, knowledge of automotive radar and git...

Autonomous driving is a challenging task and requires good scene understanding. Deep learning is very successful in many fields and radar sensors play a crucial role in scene understanding. Developing good deep learning methods for radar modalities is an important task for autonomous driving. This requires deep learning algorithms that generalize well on different domains and types of radar sensors. Develop and improve deep learning algorithms for radar! Original research, collaboration with machine learning and radars experts from academia and industry Use cutting edge methods like foundation models Publications in top-tier journals and at top-tier conferences Technical discussions and creation of new ideas within the international research team at the BCAI and collaborator Qualifications Education: excellent Master degree in Computer Science, Physics, Mathematics, Electrical Engineering or related fields Experience and Knowledge: profound knowledge on machine...

Sie arbeiten in einem interdisziplinären Forschungsprojekt mit dem Ziel die physikalisch/technischen Prozessgrundlagen für die Lasermaterialbearbeitung für zukünftige Brennstoffzellenanwendungen zu erarbeiten. Sie arbeiten dabei sowohl experimentell im Labor als auch an neuen Prozessmodellen, um die Wechselwirkung zwischen Laser, Material und Prozessergebnis zu ermitteln. Dabei ermitteln Sie die physikalischen Grenze der Prozesse, erarbeiten Wirkzusammenhänge und leiten darauf aufbauend neue innovative Prozessstrategien ab. Qualifikationen Ausbildung : exzellentes Studium im Bereich Physik, Ingenieurs-, Naturwissenschaften oder vergleichbar Erfahrungen und Know-how : Erfahrungen und Kenntnisse in der Lasermaterialbearbeitung (vorteilhaft im Bereich Laserabtrag, Laserbohren oder vergleichbare Verfahren) Erfahrungen im wissenschaftlichen Arbeiten und Aufbereiten von Forschungsergebnissen ist erwünscht Grundkenntnisse im Bereich Fertigungstechnik, im Qualitätsmanagement...

Electric vehicles deliver a major contribution to the reduction of CO2 emission and also offer a fascinating driving experience. This is defined by high power at simultaneously low energy consumption of electric vehicles and is enabled to a great extent by the semiconductor devices used in the electrical traction drive. Wide-bandgap semiconductors such as silicon carbide (SiC), gallium nitride (GaN) and gallium oxide (Ga2O3) enable the fabrication of extremely energy efficient power electronics. Become a part of the team and advance research on wide-bandgap semiconductor devices. You will design, simulate and fabricate novel semiconductor devices. Furthermore, you evaluate as well as optimize novel device concepts and create knowledge on the correlation of fabrication steps and electrical characteristics. Last but not least you will characterize devices electrically. Qualifications Education: excellent Master’s degree in Physics, Electrical Engineering or similar with...

We conduct research on state-of-the-art deep generative models that are used to enable real-world Bosch systems to be data-efficient. We are looking for a PhD student who is interested in researching creative applications of generative models (e.g. stable diffusion) as a controllable dataset representation for training and validating networks for downstream tasks. Not all data points in a dataset are equally important for the performance of a neural network. As training progresses, loss on some data points might become uninformative since the network already learned what it can from it. As such, it can be advantageous to observe the network training to serve it the right type of data at the right time. However, simply selecting data from a fixed dataset could be problematic when no image with the precise mix of attributes exists. The goal of this PhD project is to develop new learning algorithms for generating relevent data "on demand" in response to the need of the target...

Agile robot motion skills are indispensable for deploying robots in diverse and unstructured environments. The intricate nature of such skills renders manual coding impractical, thereby paving the way for learning-based techniques. Nevertheless, the practical deployment of robots in real-world scenarios demands a careful consideration of data efficiency, safety, and swift adaptability when devising motion learning methods. A promising hybrid approach entails leveraging domain knowledge, such as data geometric structures or physics models, as a powerful inductive bias. Notably, differential geometry has recently emerged as a remarkably valuable inductive bias in human motion modeling, robot motion learning, and direct policy search. This thesis will address data efficiency, safety, and adaptation for robot motion skills by investigating hybrid models that build on differential geometry and physics models. The developed algorithms will be mainly evaluated in (dual-arm) manipulation...

When designing innovative control applications, functional control performance can be traded-off with resource demand in terms of computation and communication. In dynamic distributed systems, the required control performance as well as the available system resources may change during operation. The goal of this dissertation is the extension of advanced control techniques, e.g., Model Predictive Control, to cope with distributed systems where the available computational resources are limited and time-varying. To handle such scenarios, novel adaptation techniques must be developed to trade-off between resource availability and closed-loop control performance. Furthermore, orchestration approaches shall be considered, trading resources between multiple applications while ensuring the individual control performance requirements. You will be able to evaluate your developed concepts in concrete industrial settings (e.g., autonomous driving, production systems). Furthermore, you...

An explainable and anticipating behavior of driver assistance and automated driving systems is key to their acceptance by the customers. Our research group develops novel groundbreaking behavior planning solutions using state of the art classical and machine learning algorithms. If you want to push the boundaries of the current research state of the art with us, we may have the perfect position for you. An important part of your PhD is to develop novel behavior planning algorithms that can handle interaction between the traffic agents. You apply machine learning techniques combined with model based solutions with regard to performance, training, and suitability to advance the state of the art. Furthermore, you are part of the team defining figure of merits for the reliability of the behaviour planning for automated driving. Also, you work in an agile and diverse research team with exchange across several projects. In addition, you contribute to the scientific community with...

The recent developments in the area of foundation models for perception enable a wide range of new robot applications that require an unstructured, open-world scene understanding. With these methods, robots are now able to perceive and recognize novel objects as well as their semantics. In addition, large language models show increasing generative capability that allows them to not only generate meaningful texts, but to also structure the generated text as desired. In this PhD project, the goal is to investigate how the usage of planning algorithms can be facilitated by dynamically generating problem specifications, achieved by large language models and foundation models for perception. Essentially, your work addresses the limitation of requiring a manually defined abstraction by dynamically extending the required structured planning models with newly perceived content. In addition, the combination of planning algorithms with well-known behavior and large language models...

We are looking for a motivated PhD student intereseted in high-impact cutting-edge research in Artificial Intelligence (AI) for Autonomous Driving (AD). The duration of the fully funded PhD is three years. AI provides the methodical base for constructing cognition-able machines, which can successfully support or even replace humans in various tasks. With the advent of large data sets and improved computational ressources, machine learning and in particular deep learning has become the dominating AI technology. The goal of this project is to investigate modalities to improve the generalization ability and the explainability of deep learning solutions by leveraging prior human knowledge in the form of e.g., invariances but also causal reasoning, for solving the planning task in AD. Qualifications Education: Master's degree in Computer Science, Mathematics, Physics or Engineering Experience and Knowledge: very good computer-science skills as well  as a solid mathematical...

Du hast Interesse nach Deinem erfolgreich abgeschlossenen Studium im Rahmen einer Doktorarbeit an nachhaltigen Lösungen in der Kunststofftechnik zu forschen? Dann bewirb Dich jetzt und werde im Rahmen des Bosch-Doktorandenprogramms Teil in unserem Team. Um Produkte aus Kunststoff nachhaltiger zu gestalten, muss u.a. der CO2-Fußabdruck der Werkstoffe reduziert werden. Dabei bietet der Einsatz von mechanischen Rezyklaten ein großes Potential und gilt als Schlüsseltechnologie. Insbesondere bei technischen Kunststoffen (z.B. PA66) ist jedoch die Verfügbarkeit von so genannten Post-Consumer-Rezyklaten eine Herausforderung. Da technische Kunststoffe in großen Mengen im Automobilbau eingesetzt werden, soll im Rahmen der Doktorarbeit an Closed-Loop-Konzepten für diese Werkstoffe geforscht werden. Im Rahmen Deiner Doktorarbeit wirst Du in Kooperation mit unseren produktverantwortlichen Geschäftsbereichen Bosch-Produkte aus Altfahrzeugen auswählen und beschaffen. Du untersuchst den...

Deep Learning is one of the key enabling technologies for autonomous driving, where radar is known to be a crucial input modality as it is robust for difficult whether conditions. Moreover, it can provide information about the environment that cameras cannot reach. Combining deep learning technologies with radar modality is therefore vital for the development of reliable systems. Your role will be to focus on the development and implementation of cutting-edge deep learning models focused on radar modality. Specifically, during the PhD you should conduct your own research in collaboration with machine learning and radar experts from academia as well as Bosch and publish the results in top-tier journals. You investigate new architectures by combining grid-based and point-based approaches in such a way that they have low complexity and the geometric properties of the scenes can be used as prior knowledge. Last but not least, you model temporal information for downstream tasks...

Artificial Intelligence (AI) is one of the central pillars of Industry 4.0 and Internet of Things in the global trend towards smart and highly-digitalized factories. An important approach here is to use symbolic AI, including ontologies or semantic Digital Twins, to represent experts’ knowledge, physical equipment and processes as unified conceptual Semantic Models. They in turn allow to integrate heterogeneous industrial data as uniform Knowledge Graphs. On top of ontologies and Knowledge Graphs, one can develop a wide range of AI solutions for production simulation, process monitoring, analytics and optimization. This can be achieved by combining symbolic reasoning and numeric AI methods, like Machine Learning, in hybrid AI solutions. Bosch Center for AI is active in developing such novel hybrid AI methods and systems and also in applying them in Bosch factories. The purpose of this PhD research project is three-fold. First, the PhD candidate will be looking into the problem of...

In all Bosch domains there is a trend from traditional non-connected cyber-physical system (CPS) applications (e.g. engine management) over cloud enhanced CPS that use networking and cloud computing for non-critical functional extensions (e.g. predictive maintenance) to dynamic distributed CPS where time and safety critical parts of the functionality might get shifted to the (edge) cloud (e.g. infrastructure-assisted automated driving or the Industrial Internet of Things (IIoT). Your tasks will be: Advance and apply academic research results from real-time, networking and distributed systems Research on mechanisms and architectures to constructively guarantee end-to-end real-time requirements of dynamic distributed applications Demonstrate your results in the context of Bosch applications together with an interdisciplinary team of researchers at the Corporate Research Campus at Renningen, Germany Qualifications Education: excellent Master of Science in Computer Science,...