Diedrich, Alexander

This paper focuses on the development and implementation of a diagnosis toolchain, to identify faults and recommend actions for the system operators of the environmental control and life support system of the COLUMBUS module on the International Space Station. We present a comprehensive framework which uses different aspects of artificial intelligence to efficiently identify the necessary interventions for the system operator to stabilize the system in case of emergencies and defects. Methods such as machine learning and statistical analysis, based on time-series, are used for anomaly detection to identify potentially critical situations early and issue the corresponding warnings. Diagnostic functionality enables the identification of the causes of anomalies, integrating expert knowledge and pattern recognition algorithms to achieve accurate diagnostic results. The localization of affected system parts is crucial as fault propagation can obscure the origin of anomalies. A vital aspect of the AI system is determining possible reconfiguration measures according to the behavior of the system, offering operators various operational continuance variants in the event of damage. Based on the diagnostic results, the system identifies suitable reconfiguration measures to restore normal operation or minimize potential damage. An additional supervision module based on qualitative system models is then used to monitor, evaluate, and assess the effects of these interventions. An MLOps platform facilitates the seamless integration of the framework into existing processes, providing an agile solution for fast and reliable development, scaling, and standardized integration interfaces. The successful integration of the AI toolchain at Airbus Defense and Space exemplifies this implementation’s effectiveness, significantly reducing development times and enhancing the process’s reliability and efficiency.

This report summarizes the workshop on “Learning Approaches for Hybrid Dynamical Systems”, held at the 2025 Conference on Machine Learning for Cyber-Physical Systems (ML4CPS). The workshop aimed to strengthen collaboration and foster exchange between institutions engaged in research on model learning methods for hybrid CPSs. The participating research groups approach the topic from diverse perspectives, for example, from an application perspective, from a tool perspective, or from a fundamental and formal perspective. Accordingly, this paper synthesizes the discussions from the workshop and presents an overview of key perspectives on several central topics, including the taxonomy of hybrid systems, current learning paradigms and techniques, and particularly representative use cases.

Cyber-physical systems that integrate machine learning (ML)-based services and methods from the broader field of Artificial Intelligence (AI) rely on a virtual representation of the underlying real physical system. Unfortunately, depending on respective solution approaches, usually similar but rarely the same virtual representation of the physical system is required. Thus, two solutions for the same problem might use different virtual representations. Informed Machine Learning is one technique to integrate expert knowledge into AI applications. It uses techniques to combine an often proprietary and expert-defined virtual representation with data from a real cyber-physical system. But methods for Informed ML have a much higher demand on the virtual representation than, for example, traditional distance-based methods in Machine Learning. Informed ML requires domain specific knowledge, which needs to be represented in some standardized Digital Twin as its virtual representation. Practitioners benefit through some categorization indicating which Digital Twin can be used to acquire a unique virtual representation of a cyber-physical system. Especially, by using a common standardized application programming interface (API). In short: a standardized Digital Twin is needed for AI-based solutions. In this chapter, such an API for Digital Twins for AI solutions is presented and different levels of complexity for Digital Twins are defined. The suggested API is considered as an AI reference model and is verified by using it on several simulated and real examples from the process and manufacturing industries. Additionally, it is compared against currently ongoing research projects.

Creating models that are usable for fault diagnosis is hard. This is especially true for cyber-physical systems that are subject to architectural changes and may need to be adapted to different product variants intermittently. We therefore can no longer rely on expert-defined and static models for many systems. Instead, models need to be created more cheaply and need to adapt to different circumstances. In this article we present a novel approach to create physical models for process industry systems using multi-modal large language models (i.e ChatGPT). We present a five-step prompting approach that uses a piping and instrumentation diagram (P&ID) and natural language prompts as its input. We show that we are able to generate physical models of three systems of a well-known benchmark. We further show that we are able to diagnose faults for all of these systems by using the Fault Diagnosis Toolbox. We found that while multi-modal large language models (MLLMs) are a promising method for automated model creation, they have significant drawbacks.

This is an extended abstract of the manuscript "A Lazy Approach to Neural Numerical Planning with Control Parameters". The paper presents a lazy, hierarchical approach to tackle the challenge of planning in complex numerical domains, where the effects of actions are influenced by control parameters, and may be described by neural networks.

Industrial fault diagnosis exhibits the perennial problem of reasoning with partial and real-valued information. This is mainly due to the fact that in real-world applications, industrial systems are only instrumented insofar, as sensor information is required for their functioning. However, such instrumentation leaves out much information that would be useful for fault diagnosis. This is problematic since consistency-based fault diagnosis uses available information and computes intermediate values within a system description. These values are then used to compare expected normal behaviour to actual observed values. In the past, this was done only for Boolean circuits. Recently, satisfiability modulo non-linear arithmetic (SMT) formulations have been developed that allow the calculation of real values, instead of only Boolean ones. Leveraging those formulations, we in this article present a novel method to infer missing sensor values using an SMT system description and the notion of critical pairs. We show on a running example and also empirically that we can infer novel measurements for five process industrial systems. We conclude that, although SMT calculations accumulate some error, we can infer novel optimal measurements for all systems.

With the increasing complexity of highly automated cyber-physical systems (CPS), monitoring their behavior has become crucial. Failures in these systems can be costly, halt production, or even pose risks to human safety. Effective diagnosis depends on understanding the system's components, connections, and the influences among them, knowledge typically provided by experts. However, the shift towards self-diagnosing systems necessitates this knowledge be machine-readable and interpretable. This paper introduces a novel methodology that utilizes an ontology to encode knowledge about cyber-physical systems and systematically generate propositional logical expressions. These expressions can then be evaluated using state-of-the-art diagnostic algorithms to identify failure causes. Our methodology was validated using an established AI benchmark for diagnostics. We constructed an ontology description for the underlying cyber-physical system, deduced influences of system sensors from data, and successfully diagnosed induced failures, demonstrating the efficacy and applicability of our approach.

Autonomously detecting anomalous behavior based on system observations is a fundamental task for Cyber-Physical Systems (CPS). Due to the high system complexity and large number of subsystems in modern CPS, rule- or knowledge-based approaches for anomaly detection are more and more replaced by Machine Learning (ML) approaches which leverage historical CPS data. Typically, ML approaches learn a system model based on the CPS data and identify anomalous behavior based on the distance of the real CPS behavior to the predicted model behavior. However, most classical ML approaches for anomaly detection are monolithic, meaning a single ML model is fitted on a global CPS observation, making them frail to spurious correlations and confounders that originate on CPS subsystem level. We hence propose a modular approach toward anomaly detection in CPS, specifically a novel Modular Neural Network (MNN) architecture. Our architecture not only models the behavior of individual CPS sub-systems in individual MNN modules, but additionally models the dependencies of the CPS subsystems into the MNN architecture. Thereby, we omit confounding effects and spurious correlations, enabling us to identify and allocate anomalies within the CPS at subsystem level. We benchmark our MNN architecture against monolithic Neural Networks and MNN architectures that do not explicitly model CPS subsystem dependencies using a real-world dataset of an industrial robot with different anomalies. We show that by modeling real-world dependencies into a MNN architecture, we can improve the performance of autonomous anomaly detection in CPS.