In recent months, multiple near-miss incidents were reported due to runway incursions. Runway incursions are any occurrences where an unauthorized aircraft enters the protected area of an active runway (e.g., one on which another aircraft attempts to land), potentially creating a collision hazard. To increase safety in the future, Boeing is developing new capabilities to reduce runway incursions.

The Institute of Mechatronic Systems in Mechanical Engineering is conducting research into energy-efficient drive concepts for electric robot vehicles. To better assess the potential for future drive and transport concepts, these must be tested using simulations. In order to be able to assess the influence of a traffic environment, a traffic simulation of the city centre of Darmstadt is therefore used at the IMS. This is implemented in the SUMO simulation software. An initial approach to parameterizing the simulation was developed in previous works and is now to be continued.

At the IMS, the potential for increasing the efficiency of powertrains by knowing the speed to be expected in the next few seconds is being investigated on the basis of AI-based predictions.

The Kolmogorov-Arnold networks (KAN) approach presented in April 2024 promises faster training, improved accuracy and interpretability compared to conventional multi-layer perceptrons (MLP). So far, only a simple Matlab implementation of KANs with basic functionality exists.

The aim of this advertised work is to realise a KAN implementation against the background of the Matlab Deep Learning Toolbox and to use it for speed prediction.

A data set comprising over 29,000 kilometres is available as a data basis.

In order to make the lecture more exciting and flexible, we are going one step further: With new Python content and a 3D printing system for latte art developed by you, the lecture will be expanded in the future. Your task is to first build the basic function of a 3D printing system and then implement interesting functions: Voice control, two-factor authentication or a specially developed app – all to print drinkable works of art on milk foam. The ultimate goal? To create a fully functional prototype and serve your supervisor a cup of coffee they've never experienced before.

At the IMS, the potential for increasing the efficiency of powertrains by knowing the speed to be expected in the next few seconds is being investigated on the basis of AI-based predictions.

The Kolmogorov-Arnold networks (KAN) approach presented in April 2024 promises faster training, improved accuracy and interpretability compared to conventional multi-layer perceptrons (MLP). So far, only a simple Matlab implementation of KANs with basic functionality exists.

The aim of this advertised work is to realise a KAN implementation against the background of the Matlab Deep Learning Toolbox and to use it for speed prediction.

A data set comprising over 29,000 kilometres is available as a data basis.

In order to make the products of the future more sustainable, they will inevitably have to become lighter and lighter. However, this goes hand in hand with a greater susceptibility to vibration excitation and a corresponding increase in sound radiation. The challenge is to resolve this conflict through targeted design optimizations. Usually, topology optimization approaches are used here, such as gradient-based approaches or genetic algorithms.

In this work, a novel approach using reinforcement learning (RL) will be tested instead. RL is an area of artificial intelligence that is still little used in engineering and is based on one or more agents, e.g. based on neural networks, which learn to achieve a goal with rewards based on positive or negative system behavior. Using the example of an oil pan from a vehicle, the vibroacoustic behaviour is to be optimized while minimizing its weight.

The ramp-up of e-mobility brings with it challenges for integration into the electricity grid. To test grid-friendly charging, an intelligent charging station is to be developed at the IMS.

Wildfire is a common natural disaster, which causes huge damage to the environment and humans’ life and properties. To extinguish wildfires and/or minimize its impact, a good prediction of their spreading properties, such as direction and velocity, are crucial. Since wildfire spreading is a very complicated large-scale phenomenon involving complex geometries, flow conditions, and combustion, the investigation of them usually remains at a macro level without digging into the physics of some essential components of the problem.

Therefore, there is a significant gap of fundamental knowledge that hurdles a reliable simulation/forecast of wildfire spreading. To improve our understanding and thus better prediction of wildfire spreading, the fundamental combustion properties of wood, one of the essential components, need to be studied under a well-controlled condition. Under co-flow conditions, where the spreading direction of the wildfire is the same as that of wind, the ignition and combustion of trees occur in a high-temperature and low-oxygen atmosphere.

This project will apply advanced optical techniques, e.g., infrared thermometry, to study the fundamental ignition and combustion properties of woods (in the form of stick) under a representative co-flow condition that can be produced by a laminar flow reactor.

Within the research project FlowForLife, a microfluidic supply network for 3D cell clusters is being developed. Therefore, micro channels are studied with regard to their flow as well as the oxygen transport into the surrounding medium or tissue. The oxygen transport of enriched fluid into the surrounding hydrogel matrix will be measured using phosphorescent particles.