Projects
Current research projects
Here you will find publicly funded projects in which the Chair of Mechatronics is involved.
The washing process in a car wash cannot be comprehensively modeled using today's simulation methods and tools. As a result, optimizations of the car wash systems and washing programs are developed and validated in real tests with up to a hundred different vehicles, which corresponds to an effort in the range of weeks. NACSIM (Neural accelerated Carwash Simulation) aims to comprehensively model the washing process and simulate it with high performance in order to be able to replace and thus reduce previously necessary, resource-, cost- and time-intensive real tests with simulations. The basis for this is the combination of physical relationships and corresponding models with machine learning methods.
The European Climate Law aims to cut net greenhouse gas emissions by at least 55% by 2030 and make Europe's economy and society climate-neutral by 2050. Following this, there's a pressing need to exploit energy-saving opportunities that were previously overlooked due to low energy costs. Effective investments in energy, building, and production infrastructure, for example large-scale green hydrogen production, require optimizing usage of large-scale facilities' across full day cycles throughout the year. Currently, only highly simplified models exist to analyze these large-scale scenarios, which are often insufficient for achieving the set objectives.
To tackle these challenges, core standards and established modeling and simulation tools must be enhanced to scale with size, handle increasing complexity, and provide more adaptable solutions.?
This need extends to large-scale systems (LSS) and distributed controllers in the edge-cloud continuum, which demand improved runtime scalability.
OpenSCALING aims to extend the open standards Modelica, FMI, eFMI, SSP, and related toolchains to meet these challenges and support virtual engineering and the operation of future sustainable systems. Our chair focuses on research on Physics-enhanced Neural ODEs (PeN-ODEs) to support these goals.
Several industrial demonstrators will showcase how the OpenSCALING innovations are applied in the energy, building, aviation, and automotive domains through green hydrogen production, more efficient heat pumps, fuel cell propulsion, and electrified vehicles.
The AI Production Network Augsburg is an association of the 伟德国际_伟德国际1946$娱乐app游戏 of Augsburg, the Augsburg 伟德国际_伟德国际1946$娱乐app游戏 of Applied Sciences, the Fraunhofer IGCV and the DLR ZLP, which is co-financed by the Bavarian State Ministry of Science and the Arts and the Bavarian State Ministry of Economic Affairs, Regional Development and Energy as part of the Hightech Agenda Bavaria. In the AI Production Network, we are committed to networking industry with research and associations and thus creating more awareness for state-of-the-art technologies in the fields of AI and production technology.
?
Through our large network, we acquire new regional research projects with high application potential. Ultimately, this not only strengthens the portfolio of the partners involved, but also the Augsburg and Swabia region as a business and research location.
?
As part of the AI production network, AuxMe deals with the topics of AI, digital twins, network infrastructure, and collaborative robotics. We are centrally represented in the network by i.a. two senior employees and can be found at various network events.
The development efforts required for new vehicle concepts are becoming increasingly unsustainable in light of the megatrends of electrification and digitalization, along with the emergence of new competitors in the automotive sector. Diffus3D aims to further develop AI-based generative 3D models in a domain-specific manner so that technical requirements are taken into account during the vehicle model generation process. The vision of Diffus3D is the creation of 3D vehicle models at the push of a button, which meet technical requirements such as necessary installation spaces.
Many people with disabilities are unable to participate sufficiently in the working life. This is due to the fact that workplaces are not inclusive enough to compensate for the many different types of performance changes and participation restrictions. Collaborative robots are an important tool for overcoming this challenge. However, their programming today is primarily static and does not enable dynamic and autonomous interaction with humans. This makes interaction optimized on human needs impossible.
We are researching methods of human-robot teaming that enable the machine to develop a higher degree of autonomy in cooperation with humans. People with disabilities are usually subject to fluctuating abilities that make it difficult to optimize interaction with the robot. We are, therefore, researching hybrid AI approaches that combine occupational health knowledge and machine learning to quantify people's abilities. We then use these as a basis to synthesize robot behavior that supports people as optimally as possible in their everyday work. However, the robot is by no means patronizing, but rather humans should perform as much as they are capable of. This promotes acceptance and can also help related fields, such as rehabilitation, to enable people to participate more in society.
Completed Research Projects
Here you will find completed publicly funded projects in which the Chair of Mechatronics was involved.
?
With a total of 31 partners from 7 countries, UPSIM (Unleash Potentials in Simulation) creates the conditions for companies to implement simulation and the associated AI methods as a key strategic capability in the development network. In particular, UPSIM aims to use AI-supported simulation to ensure product quality and in certification.
?
UPSIM?provides the following modules for this purpose:
- Modeling and simulation reference processes and a metric for determining the readiness level of digital twins,
- Cooperation patterns for the efficient development of digital twins,
- Hybrid simulations enriched with artificial intelligence to ensure the convergence of simulation and reality, and finally
- an infrastructure for the “chained” identification of credible artifacts of digital twin simulations.
The vehicle of the future is ?smart“. The ability of a vehicle to respond to a changing environment and changing constraints by adopting its behavior in an optimal way will be considered a commodity feature. Realizing such a flexible behavior in a vehicle requires a high degree of ?self-awareness“, in other words the ability to predict the impact of its interaction with the environment. Creating models to describe the vehicle itself and its environment properly in terms of the best trade-off between fidelity and runtime performance in a short period of time and in a very cost effective way is a key success factor.
?
Classical model-based approaches are typically associated with high development efforts. Advances in the field of artificial intelligence open up new opportunities but depend on large amounts of data, besides other risks to reach a high confidence in the model.
In this project hybrid (data and physics-based) approaches shall be evaluated in concrete applications, aiming to incorporate existing physical knowledge in order to generate scalable “Proper Models” in a very data efficient way. These methods will enable the development and realization of competitive product properties and innovative new functionality for smart vehicles in siginificantly shorter time.
For more information: [https://phymos.de/](https://phymos.de/)
?
?