Projects of Group Wotawa

Global environmental issues, social inequality and geopolitical changes will pose numerous problems for our society in the future. To face these new challenges and deal with them, there is a need to understand and appropriately utilize new digital technologies such as artificial intelligence (AI), the Internet of Things (IoT), robotics and biotechnologies. The use of such new digital technologies contributes to a higher degree of digitalization, while also allowing to respond to emergency situations in a sustainable and effective way. This in turn ensures not only civil and societal safety and security, but also improves the operational readiness and efficiency in safety critical domains such as the Search&Rescue (SAR). A- A-IQ Ready drives the digital era towards Society 5.0 by promoting civil safety, digital health and co-existence between humans and AI. A-IQ Ready presents methods for localization of so-called SAR platforms which are able to navigate in tunnel scenarios without GPS signals. Using a quantum sensor, geomagnetic field measurements and appropriate sensor fusion algorithms, the position of the SAR platform is determined. In addition, new methods will be developed and demonstrated in A-IQ Ready, which allow for a highly precise assessment of both the health and alertness status of individuals in high-risk situations, as would be in the case of a SAR mission. Using modern contact and non-contact sensor technology, it will be possible to capture the cardiorespiratory status in real time, which directly correlates to the individual`s level of fatigue. This represents an innovative paradigm shift from drowsiness monitoring to sleep prediction. B- A-IQ Ready proposes cutting-edge quantum sensing, edge continuum orchestration of AI and distributed collaborative intelligence technologies to implement the vision of intelligent and autonomous ECS for the digital age. Quantum magnetic flux and gyro sensors enable highest sensitivity and accuracy without any need for calibration, offer unmatched properties when used in combination with a magnetic field map. Such a localization system will enhance the timing and accuracy of the autonomous agents and will reduce false alarms or misinformation by means of AI and multi-agent system concepts. As a priority, the communication guidance and decision making of groups of agents need to be based on cutting-edge technologies. Edge continuum orchestration of AI will allow decentralizing the development of applications, while ensuring an optimal use of the available resources. Combined with the quantum sensors, the edge continuum will be equipped with innovative, multi-physical capabilities to sense the environment, generating “slim” but accurate measurements. Distributed intelligence will enable emergent behavior and massive collaboration of multiple agents towards a common goal. By exploring the synergies of these cutting-edge technologies through civil safety and security, digital health, smart logistics for supply chains and propulsion use cases, A-IQ Ready will provide the basis for the digital society in Europe based on values, moving towards the ideal of Society 5.0.

Buildings account for a significant portion of energy consumption in the EU and Austria, estimated at 40% and 25%, respectively. Studies have shown that up to 50% of heat pumps operate at only 70-80% efficiency due to a lack ofmaintenance activities. By detecting, diagnosing, and repairing faults in heat pumps early on, energy loss can bereduced by approximately 40%. Utilizing smart energy services, such as diagnosis, can enhance energy efficiencyand reduce CO2 emissions. Further, diagnosis and repair are crucial components of maintenance that significantlyextend the lifetime of products. There are generally two AI diagnosis avenues: knowledge-based, e.g., model-baseddiagnosis (MBD), and data-driven, e.g., machine learning (ML), approaches. While each research direction has beenapplied to industrial applications, it becomes crucial to adopt a holistic approach that combines thesemethodologies to leverage the strengths of both methods, offering a more robust and comprehensive solution. Inthe project ALFA, we want to foster synergies between model-based and machine learning techniques in the contextof fault diagnosis of building automation sector. Based on a set of real-world benchmarks, we will first develop aseparate MBD and ML approach in the application domain. Second, we will focus on ways to integrate the methodsunique strengths and capabilities. There is generally a lack of data for a newly commissioned building known as thecold start problem. To mitigate this issue, we will exploit simulations based on the MBD model to generate asynthetic diagnostic data set for training the ML model. Additionally, we emphasize the interpretability of MLdiagnostics via Explainable AI (XAI), promoting transparency in decision-making processes. Finally, we will explorefurther synergy opportunities (e.g., incorporating probabilities into MBD based on data from the ML diagnosticprocess) to set new standards for the accuracy and efficiency of diagnostics in building automation.

TUG will provide research work on diagnosis, predictive maintenance, and lifetime estimation. In any case, obtaining knowledge from available data, which includes real-world and simulation data, is of importance. In this project, we will extend previous work in diagnosis utilizing models substantially. Instead of considering hand-crafted models for diagnosis and the prediction of faults, we are considering learning models from available data. In contrast to existing work, where there is often a need for having data that corresponds to the correct and the faulty behaviour, we want to obtain a model solely from correct behaviour that can be used in a similar way for model-based diagnosis avoiding the use of known faulty behaviour. In this way, classical restrictions regarding the use of model learning and other types of machine learning can be avoided. In particular, we expect to use the learned model of the correct behaviour for fault detection and localization directly, not needing information about the faulty behaviour represented in the data. In the project, TUG will provide the foundations behind the model-learning approach and also apply it to SC2, where the focus is on diagnostics during operation considering powertrains of vehicles. The research will be carried out until reaching a technology readiness level of 4 comprising experimental proofs of the concepts using parts of the powertrain and technology to be validated in the lab.

In order to strengthen European aviation industry for the future and to increase its competitiveness the European Commission released its vision for aviation Flightpath 2050 in 2011. Among other goals, it aims at the reduction of CO2 emissions by 75 % compared to 2000. In order to achieve this goal the efficiency of modern aero-engines has to be improved considerably, whereas artificial intelligence (AI) and digitalization will play a key role (BMK, 2020). The Institute for Thermal Turbomachinery and Machine Dynamics at Graz University of Technology has been investigating the aerodynamics of intermediate turbine ducts, a key component of modern aero-engines, for many years. This research provides the institute with a large and well evaluated data basis. It shall be used for AI application in the project ARIADNE. Together with an informatics institute and two Austrian SMEs following goals shall be pursued to provide tools for the optimization of future intermediate turbine ducts in aero-engines: • Setup of a data bank of the aeronautics of intermediate turbine ducts, based on measurements and simulation of different designs at various inflow conditions. The structure of the data bank shall allow a fast and efficient utilization for AI application. • Development of methods for data reduction for efficient AI application based on POD methods and Machine Learning • Development of a method for the fast flow prediction of new designs observing the physics of fluid mechanics • Development of a tool for the evaluation of measurements in turbine ducts in order to find possible sensor errors • Development of a tool for the evaluation of flow simulations of turbine ducts in order to find possible model errors or computational mesh problems • Application of the developed tools to obtain innovative knowledge of principles in the flow of intermediate turbine ducts • Finally, the developed tools shall be combined with an optimizer with the goal of fast and efficient design optimization, much faster than with flow simulation based optimizing methods

Due to an increasing demand for reducing energy consumption and increasing product utilization, there is also an increasing need for diagnostic services. In the Ph.D. proposal AI-based Diagnosis for Energy Transition and the Circular Economy (AID4ETCE), we want to tackle this challenge and provide means allowing users in energy transition and circular economy application areas to select appropriate diagnosis methods for particular products and services. We consider different diagnosis methods ranging from knowledge-based to data-driven approaches. Whereas the former ones are suitable in cases where no data is available, i.e., immediately after deployment of products and services, the latter consider data of products and services observed during operation and are more adapted to particular instances. However, in any case, no knowledge is available that compares those approaches, considering different metrics to allow an informed decision for application. Moreover, combining the diagnosis approaches seems to be a valuable add-on due to the – at least partial – complementarity of the diagnosis approaches. Such a combination would allow knowledge-based approaches to adapt to particular instances of products and services for diagnosis and data-driven approaches to enhance their explanatory capabilities. Moreover, it is worth mentioning that such a combination contributes to the open research question of how to bring together machine learning and logic-based reasoning approaches in a unique framework.

Cynergy4MIE responds to the urgency of global competition, particularly in mobility and complex product domains like human-collaborating robots. Europe's competitiveness depends on software-centric approaches, common components, and shared tools across mobility, infrastructure, and energy domains. The Cynergy4MIE project addresses the need for new approaches in converging ecosystems, focusing on software complexity, trustworthiness, and software composability. AI-assisted tools will enhance software engineering efficiency, harmonizing methodologies and components. TUG's Institute for Software Technology contributes to SC4 Emergent ADS (Automated Driving Systems). Demonstrators tackled within SC4 include traffic flow optimization and unusual traffic encounters. In the former, TUG provides expertise in optimization using AI methodologies. In the latter, TUG contributes to identifying critical scenarios based on formalized knowledge, regulations, and standards. TUG considers its own previous research work and substantially extends it to fit the needs and challenges of the Cynergy4MIE project.

In a bid to harness cutting-edge AI advancements, Slovenia and Austria, AI research frontrunners, united with two Widening countries, North Macedonia and Serbia, to propose a transformative mission to bridge the knowledge gap, and to propagate Generative AI expertise with the goal to meld AI into pragmatic medical contexts. Together, they are laying a foundation for a cross-border collaboration methodology that through training missions, workshops, seminars, and collaborative deep dive research would enrich minds and redefine medical practices. The collective effort, wisdom, and adaptability as a guiding principle, will dive into the core of the Large Language Models to explore the capability of transforming complex medical knowledge into valuable medical diagnoses and treatment recommendations. Through simulated real-world medical challenges, the team will resonate in the practicality of Generative AI and the specific requirements that such AI-based software would require to be implemented practically in the hospitals. Through this project the partners from the Widening countries will master coordination skills, while the research on Generative AI will alter the medical landscapes.

Bilateral AI provides the means to develop the foundations of “broad AI”, a new level of AI with considerably enhanced and broader capabilities for skill acquisition and problem solving, by combining sub-symbolic AI (machine learning, ML) with symbolic AI (knowledge representation and reasoning, KRR). Bilateral AI will leverage the synergy from combining these AI fields.

In Austria, the building sector requires 25 % of the total final energy demand amounting to 316 TWh in 2019. Currently Heat Pumps (HPs) generate 6 % of the energy in the building sector (6.3 TWh or 478000 t CO2eq) and over 34 TWh are still provided by non-renewable sources Invalid source specified.. Austria targets to replace oil, gas, and coal in heating systems with renewable heat sources until 2040. HPs will play a decisive role in the future of heating systems for their low carbon footprint and ability to provide both heating and cooling. Assuming HPs with a reasonable average COP of 3.5 will replace 70 % of all existing non-renewable heating systems until 2040, their projected total electricity consumption only in buildings will be about 9 TWh/a with CO2eq emissions of about 2x106 t/a. Fault detection and intelligent control measures enable energy savings at least 10 % in HPs Invalid source specified.Invalid source specified.. This would result in a potential for CO2eq savings of 200000 t/a due to methods to be developed in this project.

The Robot Soccer World Cup (known as the RoboCup) Games and Conferences are a series of competitions and events designed to promote the full integration of AI and robotics research. Robotic soccer provides a good test-bed for evaluation of various research, e.g. artificial intelligence, robotics, image processing, system engineering and multi-agent systems. In the Middle Size League (MSL) teams of fully autonomous robots with a size of up to 50cm x 50cm x 80 cm play soccer against each other. The MSL provides a serious challenge for many research disciplines including multi-robot cooperative teams, autonomous navigation, sensor fusion, vision-based perception, automatic reasoning, and mechanical design, to name only a few. All these topics have to be tackled in order to solve the RoboCup challenge. Therefore, RoboCup needs truly interdisciplinary research. Furthermore, approaches developed in the MSL will find their way to applications in other domains like service robots. As mentioned above research in the field of autonomous mobile robots is a very interdisciplinary and wide area. Therefore, hardly any group is able to achieve high quality research in all topics. Our group concentrates its work on flexible, symbol-based and robust approaches for the control of autonomous mobile robots in a wide area of domains and for various tasks. We subsume this under the name "Robust Intelligent Control for Autonomous Systems". The main research topics of our group are: robust abstarct control for mobile robots, model-based diagnosis for autonomous systems and sensor-based navigation.