Current & past projects

Human-Centered AI system

As part of the project, a human-centered AI work station (HCKIS) is being developed in which various levels of abstraction of artificial intelligence are used. The aim is to enable intelligent and interactive cooperation, cooperative venture, partnership, collaboration between humans and smart networked machines at various levels. Through the targeted use of artificial intelligence (AI), a product is to be manufactured in interaction with AI, robots and humans. The AI should not only act autonomously, but also be able to react flexibly to human interactions.

The central aim of the scheme, enterprise, aim, goal is to intensify collaboration, cooperation in the field (of) networked artificial intelligence and at the same time to create interdisciplinary points of contact within a process. AI fulfills several tasks here. An LLM model acts as an independent agent that intervenes when necessary and can control the systems connected to the work station.

Specifically, as part of the development of a shared technology platform, a scenario is being created to test new innovative ideas for future training. The 360° AI work station serves as a demonstration platform for AI methods that build on one another, particularly in the field (of) real-time-based process analysis and quality assurance. 360° here means that humans are fully assisted by AI systems during a production order. The aim is to test innovative concepts that enable more precise, efficient and adaptive control of production and manufacturing processes. In this way, the project not only makes an important contribution to the further development of smart production systems, but also opens up new perspectives for human-machine interaction in an increasingly digitalized working world.

HCKIS 2 One Pager359 KBPDF

AI for Sustainable Plastics - enabling circularity

Stricter requirements in various regulations, such as the Commercial Waste Ordinance (GewAbfV), the Packaging Act (VerpackG) and the End-of-Life Vehicles Ordinance (AltfahrzeugV), aim to reduce the amount of waste incinerated and increase material recycling rates. For plastic packaging, which currently accounts for almost 60% of the total amount of municipal waste, the new Packaging Act provides for at least 55% to be utilized; utilization by 2030. However, for economic reasons, 53% of lightweight packaging (LDPE) waste is currently sent to waste incineration plants or used as substitute fuel in cement plants for valid reasons. Recovered, sorted plastic bales are often defective. The resulting low-quality, gray, smelly regranulate is mainly used in the construction and agricultural industries. The specific aim of KI-Sustain is to use artificial intelligence (AI) to increase the usability of plastic recyclates from municipal and commercial waste by optimizing their purity in order to close loops in the plastics industry and increase recycling rates.

AI-Sustain One Pager300 KBPDF

Use of AI in planning, budgeting and forecasting

The aim of the SKIPS project is to integrate artificial intelligence (AI), in particular machine learning (ML) and neural networks (NN), into the management accounting processes of small and medium-sized enterprises (SMEs). By using external data sources such as market trends and economic data, the aim is to significantly improve forecasting accuracy. The innovative approach makes it possible to use internal data efficiently and recognize complex correlations. This increases the efficiency of planning and budgeting processes, which leads to time and cost savings and sustainably improves decision-making in SMEs.

SKIPS One Pager369 KBPDF

Interpretable anomaly detection for Sustainability in Energy, Water and Circular Economy

The early identification of unusual operating patterns in the fields (of) Energy, Water and Circular Economy allows the maintenance of a sustainable and environmentally friendly operation of technical attachments. In practice; the selection of an AI model is not based exclusively on quality indicators, but also takes into account the traceability and thus implicitly the trustworthiness of the models. The project objective is to develop AI procedures for anomaly detection in industrial time series that are both powerful and ensure interpretability.

FiAZ 2.0 One Pager443 KBPDF

Deep Learning Super Resolution for improved mapping of the finest structures in industrial quality assurance

Technically, the process, procedure is to be realized from two modules. Firstly, an adaptation of a deep learning-based super resolution network for material microscopy and secondly, depending on the application, a specific convolutional neural network for segmentation or defect detection in materials. First, the low-resolution microscopy image is increased by an elected scaling factor (e.g. factor 4) using the super resolution network. When determining grain boundaries in copper, for example, this makes fine, poorly resolved grain boundaries clearer, which simplifies their identification.

SuReQA One Pager369 KBPDF

General multimodal inspection using generative pre-trained transformers

In industrial applications, classic Computer Vision systems are often the only practicable solution due to a lack of training data. However, their development is time-consuming and cost-intensive, as it requires in-depth expertise and many communication loops between programmers and domain experts. At the same time, there is a growing need for flexible, visually customizable image processing solutions - especially for small and medium-sized companies. AMI aims to close this gap: Using multimodal AI that combines text and images, comprehensible and editable image processing workflows are automatically generated from just a few examples. This makes it possible for the first time to efficiently solve and flexibly adapt complex visual inspection tasks without programming knowledge - for accessible, explainable and certifiable industrial image processing.

AMI One Pager274 KBPDF

Development of self-learning systems in an industrial environment

Machine learning systems (ML systems) used in industrial applications for quality control are usually customized for the respective application. This implies that the training data must be collected specifically for the respective problem and, in particular, carefully labeled. This process is both time-consuming and costly. The aim is to create a system that contains algorithms and models to simplify this process. The aim is to ensure that recurring, similar and at the same time complex analysis tasks can be handled automatically and with as little effort as possible. The detection of anomalies, outliers or deviations helps to improve the process by making it much easier to label the learning data.

One Pager SeSy1.00 MBPNG

Transferable framework for interpretable anomaly detection for industrial time series data

In industrial applications, the selection of an ML model is not only based on quality metrics, but also on the interpretability, traceability and thus implicitly also trustworthiness of the models. The aim is to develop AI processes, procedures for anomaly detection in industrial time series that are more powerful and yet interpretable.

One Pager FiAZ346 KBPNG

AI-supported process stabilization for 3D printing and polishing

The aim of the KIUPRO project is to use AI methods, especially machine learning, to stabilize both 3D plastic printing and the computer-controlled polishing process on the basis of process sensor technology and thus achieve higher component quality. As a result, a machine learning model is to be made available that is able to predict quality parameters for different processes based on sensor data.

One Pager KIUPRO551 KBPNG

AI-supported marketplace for regionally generated renewable electricity based on dynamic price models

The aim of this research project is to develop and implement an integrated platform that uses AI to forecast consumption and generation data to best match regional/local demand and regional/local supply of electricity from renewable energies on a marketplace using dynamic pricing models.

One Pager KI-REN Marketplace569 KBPNG

AI-supported process optimization for the recycling of post-industrial waste

The aim of the scheme, enterprise, goal is to use artificial intelligence (AI) to enable the recycling of post-industrial waste for new components to the extent that the quality requirements for a new product are met.

One Pager AI ProWitA431 KBPNG

Improvement of detection accuracy by extension of condition classes for transient airborne and structure-borne noise signals using artificially generated process data.

The aim of the TRANGAN project is to develop a generally applicable method for artificial data generation. This method should enable the creation of artificial transient process data. The impact of the solution approach is the further development of known methods for GAN data generation from quasi-stationary signals to the generation of transient signals. This significantly simplifies the development of AI-based condition monitoring systems with the requirement for high classification accuracy and shortens the development time by reducing the number of tests required.

One Pager TRANGAN412 KBPNG

Human-centered AI system

A human-centered AI system [HCKIS] is to be set up, which will enable smart networked machines to interact with humans on different levels using AI systems. As part of, within the framework of the goal, a joint technology carrier is to be set up, which will be used in particular to demonstrate methods for real-time-based process analysis and quality assurance using artificial intelligence.

One Pager HCKIS389 KBPNG