The Business Process Management and the Internet of Things (BP-Meet-IoT) workshop series established itself approximately ten years ago as a platform for exploring the intersection of Business Process Management (BPM) and the Internet of Things (IoT). The workshop idea originated in a Dagstuhl seminar in 2016 and since then, the workshop has grown into a format bringing an audience to discuss a wide range of topics at the intersection of BPM and IoT. The founding of the BP-Meet-IoT workshop evolved at a time when the IoT and BPM domains themselves were mature, but its intersection remained largely unexplored. The workshop aimed to bring together the community to discuss the challenges and opportunities presented by this new paradigm. From its inception, the focus was on integrating process modeling, analysis, and automation with the specific requirements and possibilities of IoT environments. The increasing use of IoT devices, the growing complexity of interconnected systems, and the need for efficient process automation called for related research. The BPM IoT manifest [17] has become a frequently cited reference. Within the manifest that summarizes the key concepts, challenges, and future directions for combining BPM and IoT, the focus was on leveraging data from connected devices to automate, monitor, and optimize business processes.

Efficient business process management (BPM) relies on effective resource allocation. While Industry 4.0 adopts a system-centric perspective that overlooks human resources, Industry 5.0 emphasizes human-centered processes by aiming to account for work-induced physical and mental fatigue and other wellbeing indicators affecting employee health and organizational performance. However, despite these developments, current BPM technologies do not yet consider real-time wellbeing of human resources when allocating them to process activities. The aim of this study is to identify the functional requirements for a dynamic resource allocation mechanism that can enhance a BPM system to improve the real-time physical and mental states of human resources. The requirement elicitation follows a systematic approach through domain analysis and scenario analysis, the latter informed by interviews and observations from real-life scenarios. By formulating these requirements, we establish a foundation for BPM systems that can actively support and enhance the wellbeing of the human resources involved in the processes.

Business processes (BPs) that are enhanced with Internet of Things (IoT) technology, such as smart manufacturing processes, leverage IoT devices like sensors to monitor and capture contextual data from the physical environments where processes are executed. While the execution of BPs is typically recorded in information systems as event logs, IoT-enhanced BPs also produce IoT data that can offer valuable contextual insights. However, existing process mining techniques, which typically focus on the control-flow perspective, often miss key insights into the dynamic interplay of process activity sequences and IoT data—such as how certain IoT readings may trigger or affect specific process activities. To address this gap, we propose TROPICCAL, a new technique for multi-perspective trace clustering that integrates three key perspectives: the control-flow perspective, the trace attribute data perspective, and the time series (TS) sensor data perspective. The main novelty of TROPICCAL is the analysis of so-called context events as part of the TS data perspective. These events mark process-significant happenings detected in the TS sensor data. Furthermore, in order to unravel more insights from the output of our technique, we propose approaches for cluster explainability based on permutation feature importance. We demonstrate the efficacy of our approach and compare it with the most related and advanced approaches from the literature using a real-world manufacturing use case. Expert evaluation through in-depth interviews reveals that TROPICCAL offers better insights into the multi-perspective variants of the process.

The continuous flow of data collected by Internet of Things (IoT) devices, has revolutionised our ability to understand and interact with the world across various applications. However, this data must be prepared and transformed into event data before analysis can begin. In this paper, we shed light on the potential of leveraging Large Language Models (LLMs) in event abstraction and integration. Our approach aims to create event records from raw sensor readings and merge the logs from multiple IoT sources into a single event log suitable for further Process Mining applications. We demonstrate the capabilities of LLMs in event abstraction considering a case study for IoT application in elderly care and longitudinal health monitoring. The results, showing on average an accuracy of 90% in detecting high-level activities. These results highlight LLMs' promising potential in addressing event abstraction and integration challenges, effectively bridging the existing gap.

Business processes (BPs) are more and more enhanced with IoT devices, i.e., sensors monitoring relevant parameters of the physical environment and actuators automating certain tasks. Integrating the data generated by these IoT devices with typical process event data opens the door for the analysis of IoT-enhanced BPs at an unprecedented level of detail. In particular, IoT data enable the contextualisation of BPs to improve the prediction of next activities or process outcomes. However, existing predictive process monitoring techniques are not able to take IoT data as input, due to three challenges: i) the granularity gap between IoT and process data, ii) the uncertain scope of relevance of IoT data, and iii) the dynamicity of IoT data. In this paper, we examine these challenges and put forward three novel approaches for IoT-enhanced predictive process monitoring. These three approaches are evaluated in a real-life manufacturing case study analysing an IoT-enhanced production process where IoT and process event data are combined to predict when an activity of interest, prompted by specific IoT data patterns, will take place. Our evaluation shows that approaches integrating IoT data outperform traditional control-flow based techniques.

Advances in Internet-of-Things (IoT) technologies have prompted the integration of IoT devices with business processes (BPs) in many organizations across various sectors, such as manufacturing, healthcare and smart spaces. The proliferation of IoT devices leads to the generation of large amounts of IoT data providing a window on the physical context of BPs, which facilitates the discovery of new insights about BPs using process mining (PM) techniques. However, to achieve these benefits, IoT data need to be combined with traditional process (event) data, which is challenging due to the very different characteristics of IoT and process data, for instance in terms of granularity levels. Recently, several data models were proposed to integrate IoT data with process data, each focusing on different aspects of data integration based on different assumptions and requirements. This fragmentation hampers data exchange and collaboration in the field of PM, e.g., making it tedious for researchers to share data. In this paper, we present a core model synthesizing the most important features of existing data models. As the core model is based on common requirements, it greatly facilitates data sharing and collaboration in the field. A prototypical Python implementation is used to evaluate the model against various use cases and demonstrate that it satisfies these common requirements.

The increasing availability of data from diverse sources presents new opportunities for inclusive and data-driven modeling, particularly in the area of human behavior analysis. Process discovery, as a data-driven technique of Process Mining (PM), can be used to extract behavioral and workflow models from event logs. However, this technique requires well-structured event logs that accurately reflect actual processes. In many cases, discovery approaches rely on single-source datasets and rarely integrate heterogeneous data and contextual information. This lack of integration limits the ability to extract deeper insights into human behavior and hinders comprehensive and interpretable modeling. To address this, we propose a methodological approach called Human Behavior Monitoring with Unified Log and Contextualized Process Discovery (HB-UniContex). HB-UniContex consolidates data from multiple sources, including smart sensing systems, multimedia data streams and human-generated data sources, into a unified event log that supports context-aware process discovery and detailed visualization. HB-UniContex enables a clear extraction and interpretation of human behavior in relation to contextual information. We demonstrate the applicability of this method through a case study in an Ambient Assisted Living (AAL) setting, analyzing human behavior models and examining the relationship between mood states and daily activities. Our findings reveal specific behavior patterns associated with different mood states, providing precise insights into how an individual’s daily mood correlates with certain daily activities.

Analyzing business processes is important for organizations aiming to optimize operations and identify inefficiencies. Traditional discovered process models often lack sufficient contextual depth, limiting the interpretability and actionability of the revealed activity process flows. This paper addresses the challenge of balancing interpretability with complexity in discovered process models by introducing a new context-driven method, namely Contextualized Activity hieRarchies for Process dIscovery (CARPI). CARPI consists of a detailed five-step process to identify, extract, and integrate meaningful contextual variables into core activities flows in the process to enhance model clarity and decision-making support. We implement and validate this method using a real-world case study in manufacturing and the BPI Challenge 2017 dataset, demonstrating how the integration of relevant contextual variables refines process models to make activity flows more interpretable and actionable. This contribution advances the field of process mining by offering a clear and structured method to enrich process models with important context variables, laying the foundation for more insightful and effective business process management and improvement.

Process mining traditionally assumes centralized event data collection and analysis. However, modern Industrial Internet of Things systems increasingly operate over distributed, resource-constrained edge-cloud infrastructures. This paper proposes a structured approach for decentralizing process mining by enabling event data to be mined directly within the IoT systems edge-cloud continuum. We introduce ContinuumConductor a layered decision framework that guides when to perform process mining tasks such as preprocessing, correlation, and discovery centrally or decentrally. Thus, enabling privacy, responsive and resource-efficient process mining. For each step in the process mining pipeline, we analyze the trade-offs of decentralization versus centralization across these layers and propose decision criteria. We demonstrate ContinuumConductor at a real-world use-case of process optimazition in inland ports. Our contributions lay the foundation for computing-aware process mining in cyber-physical and IIoT systems.

Social sustainability has often been overlooked in business process evaluation compared to the environmental and economic dimensions. However, the concept of social sustainability, which emphasizes fair and just practices towards stakeholders and society, is gaining prominence in business, yet its definition remains ambiguous, obstructing its integration into business process management. This paper investigates social sustainability indicators in current literature related to business processes. A systematic literature review reveals a diverse set of 254 indicators used to assess social sustainability, ranging from standard indicators to context-specific ones, and finds the most common indicators in the realm of employee well-being and development and human rights and labor practices. Our research highlights the need for further studies to address the challenges of defining and measuring social sustainability in the context of business processes.

Time series anomaly detection (TSAD) finds many applications such as monitoring environmental sensors, industry KPIs, patient biomarkers, etc. A two-fold challenge for TSAD is a versatile and unsupervised model that can detect various different types of time series anomalies (spikes, discontinuities, trend shifts, etc.) without any labeled data. Modern neural networks have outstanding ability in modeling complex time series. Self-supervised models in particular tackle unsupervised TSAD by transforming the input via various augmentations to create pseudo anomalies for training. However, their performance is sensitive to the choice of augmentation, which is hard to choose in practice, while there exists no effort in the literature on data augmentation tuning for TSAD without labels. Our work aims to fill this gap. We introduce TSAP for TSA "on autoPilot", which can (self-)tune augmentation hyperparameters end-to-end. It stands on two key components: a differentiable augmentation architecture and an unsupervised validation loss to effectively assess the alignment between augmentation type and anomaly type. Case studies show TSAP's ability to effectively select the (discrete) augmentation type and associated (continuous) hyperparameters. In turn, it outperforms established baselines, including SOTA self-supervised models, on diverse TSAD tasks exhibiting different anomaly types.

More and more so-called IoT-enhanced business processes (BPs) are supported by IoT devices, which collect large amounts of data about the execution of such processes. While these data have the potential to reveal crucial insights into the execution of the BPs, the absence of a suitable event log format integrating IoT data to process data greatly hampers the realisation of this potential. In this paper, we present the Native Iot-Centric Event (NICE) log, a new event log format designed to incorporate IoT data into a process event log ensuring traceability, flexibility and limiting data loss. The new format was linked to a smart spaces data simulator to generate synthetic logs. We evaluate our format against requirements previously established for an IoT-enhanced event log format, showing that it meets all requirements, contrarily to other alternative formats. We then perform an analysis of a synthetic log to show how IoT data can easily be used to explain anomalies in the process.

The Internet of Things (IoT) plays a crucial role in applications for the emerging Industry 5.0 and Society 5.0. However, due to the complex nature of IoT systems, it is essential to have well-defined conceptual models for effective understanding. A good way to enhance this understanding is by incorporating categorization into the conceptual models, which can be achieved through the use of categorization patterns. This paper evaluates how different categorization patterns have an impact on the quality of IoT systems. Specifically, two categorization patterns (Inheritance and Type pattern) and a base model are used to create conceptual models in an IoT setting and are evaluated by the Quality Model for Object-Oriented Design (QMOOD) metric. We examine the impact of these patterns on different design properties and quality attributes, such as reusability, flexibility, understandability, functionality, extendibility, and effectiveness. Additionally, we use the MERODE framework, which allows us to automatically transform the models into code, enabling a more comprehensive evaluation of the patterns' impact through the generated code. To evaluate the patterns we use two IoT use cases: 1) employees' well-being in a smart factory (Industry 5.0), and 2) the daily behavior of solo residents in a smart home (Society 5.0). The results indicate that the Inheritance model exhibits the highest reusability and functionality in comparison to the other models, making it well-suited for complex projects with significant reuse requirements. This evaluation provides valuable and extensive insights for developers, helping them in understanding the impact of different categorization patterns in the final system.

Understanding the complex interactions between external and internal factors that influence pollinator foraging behaviour is essential to understand ecosystem functioning, design agricultural practices or develop effective conservation strategies. However, it remains challenging to collect large and reliable data sets with reasonable personnel and workload. In this study, we present a wireless and cost‐effective robotic flower equipped with internet of things (IoT) technology that automatically offers nectar to visiting insects while monitoring visitation time and duration. The robotic flower is easy to manipulate and settings such as nectar refill rates can be remotely altered, making it ideal for field settings. The system transmits data completely wirelessly and autonomously, is mobile and easy to clean. The prototype settings allow for approximately 2 weeks of uninterrupted data collection for each battery charge. As a proof‐of‐concept application, a foraging preference dual choice experiment with bumblebees was performed. On average, more than 7000 flower visits per colony were registered daily with a set‐up consisting of 16 robotic flowers. The data show a gradual preference shift away from the pre‐trained low concentration, confirming the hypothesis of favouring sugar water with higher concentration. The robotic flower provides accurate and reliable data on insect behaviour, significantly reducing the price and/or labour costs. Although primarily designed for (bumble)bees, the system could be easily adapted for other flower‐visiting insects. The robotic flower is user‐friendly and can be easily adapted to address a wide range of research questions in pollination ecology, conservation biology, biocontrol and ecotoxicology, and allows for detailed studies on how nectar traits, flower colour and shape or pollutants affect foraging behaviour.

In this paper, we present a novel case study for the application of process mining (PM) on data captured by Internet of Things (IoT) devices. In this case study, IoT sensors in robotic flowers are employed to gather a comprehensive dataset on bee colony behaviour through a series of experiments conducted under various conditions. The primary objective of this paper is to investigate the applicability of PM to analyse the collected sensor data, map bee colony behaviour, and uncover the learning patterns exhibited by the bees during these experiments. Encouraging results have emerged from this research, demonstrating the feasibility of converting the collected sensor data into event logs that can produce insights on bees foraging behaviour using a PM tool. This case study serves as a solid foundation for future research endeavours on the application of PM to similar processes that can be monitored using IoT devices.

The recent surge in foundation models for natural language processing and computer vision has fueled innovation across various domains. Inspired by this progress, we explore the potential of foundation models for time-series forecasting in smart agriculture, a field often plagued by limited data availability. Specifically, this work presents a novel application of $\texttt{TimeGPT}$, a state-of-the-art (SOTA) time-series foundation model, to predict soil water potential ($\psi_\mathrm{soil}$), a key indicator of field water status that is typically used for irrigation advice. Traditionally, this task relies on a wide array of input variables. We explore $\psi_\mathrm{soil}$'s ability to forecast $\psi_\mathrm{soil}$ in: ($i$) a zero-shot setting, ($ii$) a fine-tuned setting relying solely on historic $\psi_\mathrm{soil}$ measurements, and ($iii$) a fine-tuned setting where we also add exogenous variables to the model. We compare $\texttt{TimeGPT}$'s performance to established SOTA baseline models for forecasting $\psi_\mathrm{soil}$. Our results demonstrate that $\texttt{TimeGPT}$ achieves competitive forecasting accuracy using only historical $\psi_\mathrm{soil}$ data, highlighting its remarkable potential for agricultural applications. This research paves the way for foundation time-series models for sustainable development in agriculture by enabling forecasting tasks that were traditionally reliant on extensive data collection and domain expertise.

The IoT and Business Process Management (BPM) communities co-exist in many shared application domains, such as manufacturing and healthcare. The IoT community has a strong focus on hardware, connectivity and data; the BPM community focuses mainly on finding, controlling, and enhancing the structured interactions among the IoT devices in processes. While the field of Process Mining deals with the extraction of process models and process analytics from process event logs, the data produced by IoT sensors often is at a lower granularity than these process-level events. The fundamental questions about extracting and abstracting process-related data from streams of IoT sensor values are: (1) Which sensor values can be clustered together as part of process events?, (2) Which sensor values signify the start and end of such events?, (3) Which sensor values are related but not essential? This work proposes a framework to semi-automatically perform a set of structured steps to convert low-level IoT sensor data into higher-level process events that are suitable for process mining. The framework is meant to provide a generic sequence of abstract steps to guide the event extraction, abstraction, and correlation, with variation points for plugging in specific analysis techniques and algorithms for each step. To assess the completeness of the framework, we present a set of challenges, how they can be tackled through the framework, and an example on how to instantiate the framework in a real-world demonstration from the field of smart manufacturing. Based on this framework, future research can be conducted in a structured manner through refining and improving individual steps.

Contextualization is an important challenge in process mining. While Internet of Things (IoT) devices are collecting increasing amounts of data on the physical context in which business processes are executed, the IoT and process mining fields are still considerably disintegrated. Important concepts such as event or context are not understood in the same way, which causes confusion and hinders cooperation between the two domains. Accordingly, in the paper, a consolidated model to bridge the conceptualization gap between the IoT and process mining fields, based on IoT ontologies and business process context models, is presented. This consolidation based on an initial model was obtained after an extensive validation both with an expert panel and with case studies. The results of the expert survey show that the model properly describes the links between the IoT and process mining and that it has added value for IoT process mining. Furthermore, the model was refined according to the experts’ feedback. Accordingly, the paper’s key contribution consists of a common reference model that can instigate true interdisciplinary research connecting IoT and process mining.

In recent years, IoT sensors have enabled smart agriculture to grow rapidly with many compelling real-world applications. One such application is the case of smart irrigation. A particular interest exists in forecasting soil water potential to allow for the establishment of more efficient irrigation systems. Nonetheless, forecasting soil moisture is a complex task and depends on various information sources. Most existing work relies on local approaches, which are less effective at leveraging shared information across different data sources. Therefore, this paper presents a robust global approach for soil water potential forecasting, combining various environmental factors through the cutting-edge Temporal Fusion Transformer. Our proposed approach outperforms established baselines for forecasting soil water potential. As such, this work contributes to the growing body of research on data fusion in real-world applications.