Background: Predicting hypoglycemia in diabetes management remains a substantial challenge, especially for individuals with Type 1 Diabetes (T1D), advanced Type 2 Diabetes (T2D), and prediabetes. Existing prediction systems are largely dependent on continuous glucose monitoring (CGM) data and offer limited accuracy for extended prediction horizons. These systems often fail to account for the complex, individualized physiological variations inherent to each patient. Traditional monitoring methods are primarily invasive and lack the foresight needed for timely intervention. This limitation results in a heightened risk of severe complications and a significant decline in the quality of life (QoL) of patients with diabetes. Therefore, addressing these limitations requires an innovative, personalized, and multimodal approach to enhance the efficacy of hypoglycemia prediction and empower proactive diabetes management. Objectives: The core aim of this research is to develop and validate advanced predictive models for personalized hypoglycemia prediction through three primary domains: (i.) Methodological Development focused on advanced algorithm development, temporal modeling, and the creation of semantic frameworks to capture complex physiological interactions. (ii.) Data Integration and Analysis emphasizes multimodal data integration, the use of non-invasive monitoring approaches, and advanced pattern recognition to enhance the predictive power of the models. (iii.) the Implementation Framework aims at establishing personalization strategies, assessing clinical implementation, and optimizing technological solutions for embedding predictive models into wearable devices. Collectively, these objectives work towards an innovative, personalized, and practical approach to managing hypoglycemia in individuals with diabetes. Methods: This cumulative thesis synthesized findings from five peer-reviewed publications, utilizing data from three complementary datasets: D1NAMO (n=7, Type 1 Diabetes (T1D) patients), BIG IDEAs Lab (n=16, prediabetic individuals), and MIMIC-III (glucose-insulin paired data from 9 518 patients). Key methodologies included shapelet-based feature extraction to identify distinctive physiological patterns indicative of hypoglycemia and semantic integration using ontologies and knowledge graphs for enhanced data context. Both traditional machine learning (ML) and Deep Learning (DL) models, such as Fully Convolutional Network (FCN) and Residual Network (ResNet), were evaluated for their predictive capabilities. Model validation implemented holdout and leave-one-person-out cross-validation. This approach emphasized personalized performance, temporal alignment, and the integration of multimodal physiological signals to ensure robust, individualized hypoglycemia prediction. Results: This research resulted in several key advancements in predictive modeling for hypoglycemia: (1) The FCN achieved 97% accuracy in predicting the time-to-hypoglycemia, extending prediction horizons up to 48 hours; while the ResNet model achieved 94% accuracy, emphasizing the role of model architecture in optimizing prediction capabilities. (2) Temporal analysis revealed critical glucose normalization patterns within a 1–4 hour timeframe before hypoglycemic episodes, underscoring opportunities for preventive interventions. (3) Shapelet-based analysis revealed varying model performances: the three-layered Convolutional Neural Network (CNN) achieved 76% accuracy with heart rate data, while the two-layered CNN model reached 67% accuracy. In comparison, traditional machine learning (ML) approaches showed complementary strengths – Random Forest Classifier (RFC) demonstrated 73% accuracy with heart rate and 69% with breathing rate data, and Support Vector Machine (SVM) achieved 56% accuracy with heart rate and 65% with breathing rate data. These differences in performance demonstrated that advanced architecture optimization is vital for capturing personalized physiological responses. (4) Correlation analyses demonstrated substantial inter-individual variability in glucose-heart rate relationships, with correlation coefficients ranging from -0.4087 to 0.1882, thus highlighting the necessity for tailored modeling approaches. (5) Integration of a semantic framework, utilizing ontologies and knowledge graphs, uncovered previously undetectable patterns through structured representations of patient-specific factors. This structured knowledge representation contributed to improve interpretability and prediction capabilities. (6) Classification models with temporal pattern modeliing, adapted to patient-specific glucose fluctuations achieved accuracy rates ranging from 84% to 99% for different individuals, thus, highlighting the importance of personalization in predictive modeling. Conclusion: This research demonstrated that integrating multimodal physiological data, advanced temporal modeling, and semantic knowledge frameworks significantly enhances the prediction of hypoglycemic events. Also, by employing personalized modeling approaches, predictive accuracy per patient can be improved, enabling timely and patient-specific interventions. These advancements pave the way for transforming hypoglycemia prediction into a proactive and individualized system, ultimately contributing to better clinical outcomes and improved QoL for patients with diabetes.

This dissertation focuses on inter-individual differences in the processing of aversive and appetitive stimuli. Approaches for clinical interventions can be developed based on the associations between deviations in stimulus processing and mental health. In this dissertation, three central processes of stimulus processing were examined: the cognitive processing of emotional stimuli, biases of which are associated with the development of depression, affective learning, which is often used as a model for addictive disorders, and the reaction to acute stress stimuli as a relevant influencing factor for the development of many clinical conditions. Special consideration was given to the role of habitual anxiety coping, as a repressive coping style is associated with an increased prevalence of numerous stress-related illnesses. To investigate these factors, three elaborate studies were conducted in which various physiological measures such as electrodermal activity and cardiac activity, subjective ratings and implicit measures of cognitive biases were used. The results indicate that inter-individual differences play a role in all the stimulus processing steps considered. Study I showed that habitual anxiety coping can influence automatic action tendencies towards positive stimuli, with repressors showing an increased approach tendency. This suggests that cognitive biases may be maladaptive in principle and not only to negative stimuli. Study II showed that appetitive conditioning can lead to comparable cardiac CRs as aversive conditioning and thus provide a new peripheral physiological measure of CRs in appetitive conditioning paradigms. Furthermore, the study provided evidence for a relationship between appetitive and aversive CRs on a subjective, but not on a physiological level. In Study III, no altered stress response known from repressors in the form of a weakened subjective stress perception and an increased physiological stress response could be observed under non-social stress. This indicates that this reaction could possibly be triggered primarily by social stress. Overall, the results provide evidence that inter-individual differences may play a role in emotional stimulus processing. Implications for theory and practice are discussed.

The control of the combustion process of sewage sludge incineration plants has been very little investigated, because of the complexity of the sewage sludge material. Previous experiences and researches have shown that in the praxis most of the thermal waste treatment plants are still almost exclusively equipped with PI and PID controllers. Due to the inhomogeneity and the short-term fluctuations in the calorific value of the incinerated fuels, the required control performance cannot be achieved by using conventional control methods. Within the framework of this Work, the realization of an automatic stable operation of sewage sludge combustion in the stationary fluidized bed incinerator was investigated. Thus can make an important contribution to achieve better combustion control. A uniform and stable fluidization of the bed material are crucial for a homogeneous temperature distribution and an efficient combustion. In order to achieve this and thus to realize an optimal plant operating condition, two complex three-dimensional numerical models of the fluidized bed combustors were developed in the first part of this dissertation by using the commercial CFD software package Fluent 14.5. In addition to the numerical investigations, comparative experimental investigations were also carried out on the bubbling fluidized bed combustor. The results of the minimum fluidization velocity and the resulting pressure drop obtained from the numerical flow simulations were compared and validated with the experimental measurements. The second part of this thesis deals with the mathematical modelling and controller design for the stationary fluidized bed incinerator Salierweg Bonn. The implementation was done in Matlab (version 2014b). At the beginning, an experimental multivariable system of the bubbling fluidized bed incinerator was developed from measured input and output variables of the unit. After developing the multivariable mathematical model, the accuracy of the control systems was checked and validated by using collected operational data. After the validation, the combustion control methods were designed and tested based on these developed multivariable control systems. To optimize the control quality, the decoupling controllers, which eliminate the interaction between the controlled variables, were also considered. In order to define all possible operating states of the bubbling fluidized bed incinerator and to meet all required conditions for various process variables, the control results of the decentralized PI and an adaptive fuzzy PI controller were evaluated and compared using Matlab simulations. The main objective of the present work is the development of continuous and adaptive combustion control to the constantly changing properties of the sewage sludge fuel in order to optimize the combustion process, so that the system stability and optimum combustion for all changing operating conditions is assured.

Gerahmt von einer Paraphrase des Vortrags von Max Weber über »Wissenschaft als Beruf« gibt der Autor Rechenschaft über sein Berufsleben. Dazu zieht er eine Reihe von Texten heran, die als Stichprobe seiner Arbeiten dienen sollen: Rückblick auf die Dissertation und Folgeuntersuchungen zu August Herrmann Francke; Berichte über – meist inhaltsanalytische – Forschungsvorhaben; Beobachtungen und kritische Anmerkungen zum akademischen Unterricht; kritische Erörterungen zu seiner Wissenschaft. So beklagt er das, was er als »Ansatzologie« bezeichnet, und fragt, ob in seiner Wissenschaft noch Platz für die Kategorie des »Bösen« ist, und antwortet: »Tua res agitur« – »Auch deine Sache ist’s, wenn die Hütte des Nachbarn brennt.«