ProAmbitIon: Online Process Conformance Checking with Ambiguities Driven by the Internet of Things
Type
fundamental research project
Start Date
November 2022
End Date
October 2025
Acronym
ProAmbitIon
Status
ongoing
Keywords
Process Mining
Conformance Checking
Explainability
Internet of Things
Ambiguous Process Models
Description
The ongoing digitization of processes in all domains of everyday life driven by IT systems shows great potential for process automation, analysis and optimization. However, the digital traces of processes executed in the real world--especially those with human involvement--are often incomplete, too coarse-grained, or captured for individual process steps but not linked to the overall process context. This limits the possibilities for an automated checking of process conformance between the normative process descriptions and their actual execution.In this project--ProAmbitIon--we propose to use the Internet of Things (IoT) as enabler to close the gap between the physical world process execution and its digital representation. New software-controlled sensors and actuators emerging with the IoT promise to enable a fine-grained monitoring of activity executions and the correlation with their underlying processes. The resulting digital representations (traces) of the executed physical world processes lay the foundation for an automated online conformance checking as a process mining activity to detect deviations and non-conformance at runtime. Moreover, when considering human-centered processes, the process knowledge and descriptions are often provided in unstructured informal documents that allow multiple valid (ambiguous) interpretations and executions. Current conformance checking techniques are not capable of handling these ambiguous descriptions in online settings. They rely on a clearly specified formal process model--an assumption that we will relax within ProAmbitIon.We aim at developing new approaches for IoT-driven process conformance checking that are able to cope with ambiguities originating from informal process descriptions and lack of process-related data in detected activities. This includes a user-friendly approach to enrich process descriptions with IoT-related, pattern-based monitoring points to be provided via domain experts. Based on these monitoring points, new mechanisms for abstraction and correlation of IoT data with the execution of processes will be developed. This process execution data has to be be able to represent ambiguities that emerged during event abstraction and correlation. This data serves as input for novel conformance checking algorithms that are able to handle and resolve ambiguities in offline and online analyses. To make the results of conformance checking understandable by end-users and to interactively resolve remaining ambiguities, new concepts for providing interpretable feedback about process conformance will be developed.The project will be conducted following principles of design science developing new artifacts including constructs, frameworks, models, methods and instantiations. Requirements and evaluations will be based on real-world scenarios from healthcare and manufacturing, and will be additionally grounded in literature. The project combines the research competencies and strengths of both principal investigators in Switzerland and Mexico on the topics of process mining, conformance checking, software engineering and IoT, and is supported by renowned researchers and domain experts.With ProAmbitIon we will reduce the gap between processes executed in the physical world and their digital representations. The IoT hereby serves as a new source for process execution data, which will impact process mining in several ways. First, it relaxes the assumption that a central IT system is available to monitor and control the execution. Here the pattern-based monitoring points specified for process elements and mechanisms for event abstraction and correlation will simplify the generation of process event logs and streams based on a new framework for IoT-based event processing. Second, it is no longer necessary to have clearly specified process models for conformance checking as we also allow informal descriptions and the scope of conformance checking will be expanded to online settings where process executions are analyzed at runtime. Our conformance checking approaches will provide user-friendly interactive feedback in case the underlying processes have not been completely followed, resulting in new notions of partial conformance and interpretable feedback. This opens up many new application domains where human involvement and expertise play an important role and cannot be addressed with current process mining techniques.
Leader contributor(s)
Member contributor(s)
Partner(s)
Luciano García-Bañuelos, Department of Computer Science School of Engineering and Sciences Tecnológico de Monterrey
Funder
Range
HSG + other universities + partners
Range (De)
HSG + andere Unis + Partner
Eprints ID
248292
Funding code
208497
20 results
Filters
Settings
Now showing 1 - 10 of 20
- Some of the metrics are blocked by yourconsent settings
Item type:Publication, Generating Stream Processing Applications for IoT-based Process Event Abstraction: A Framework and ArchitectureEvents play an important role in information systems engineering to facilitate the analysis of relevant happenings in a system or business process via process mining. The Internet of Things (IoT) provides new ways of collecting execution-related events in real world-physical-systems using sensors and actuators. However, these new sources emit data at a too fine-grained level, which prevents process mining from deriving meaningful insights. We present a generic event abstraction framework to lift low-level data to higher level events. Starting with annotated IoT data, we generate stream processing applications that encode change patterns derived from the low-level data. These applications are then used for detecting events and activities at runtime within our proposed software architecture. We evaluate the approach for process executions in smart manufacturing and smart healthcare.Type:conference contributionJournal:International Workshop on Enterprise Modeling and Information Systems Architecture (EMISA) - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Activity and Sequence Detection Evaluation Metrics: A Comprehensive Tool for Event Log ComparisonNowadays, event logs are not only created by traditional information systems, but also new data sources such as the IoT are considered to derive and construct event logs. This makes it necessary to evaluate the quality of these detected event logs and their underlying detection methods by comparison with given ground truth logs. We present AquDeM, enabling the comparison of XES-based event logs to evaluate activity and sequence detection methods. AquDeM features 1) a Python library that allows for programmatic comparison of event logs featuring a comprehensive set of metrics, and 2) a web app for visual event log comparison.Type:conference contributionJournal:BPM 2024 Best Dissertation Award, Doctoral Consortium, and Demonstration & Resources Forum - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Ambiguity Detection in Business Process Descriptions: An Evidence and an Automated Approach(Springer Nature Switzerland, 2025-08-31); Type:conference contributionScopus© Citations 1 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Case-Based Activity Detection from Segmented Internet of Things DataThe use of Internet of Things (IoT) technologies drives the automation of business processes. However, such environments often lack process awareness and corresponding systems to monitor process executions. Due to its too fine-grained nature and variations, the direct use of data from IoT devices for monitoring is problematic, requiring an event abstraction step to lift the data to the business process level. This work investigates the application of Temporal Case-Based Reasoning (TCBR) as a novel experience-based approach to detect process activity executions in IoT data. The proposed TCBR approach uses activity signatures-representations of process and IoT data for an activity prototypeas a case base to classify unknown IoT time series data from a smart factory. A data flow architecture is presented that supports analysts in selecting a suitable activity prototype and evaluating its quality for activity detection. The results enable both, the development of high-quality activity detection services and the identification of improvement opportunities in IoT monitoring systems. The approach is evaluated using data produced by a smart factory. The results indicate that the TCBR methods used are very suitable for detecting activities in this IoT use case.Type:conference contributionJournal:Case-Based Reasoning Research and DevelopmentVolume:15662Scopus© Citations 4 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, A semi-automated approach to detecting process-level activities from sensor data(Elsevier, 2025-04); ; Silva Trujillo, Alejandra GuadalupeInternet of Things (IoT) technologies can be leveraged to monitor and check the conformance of business process executions in the absence of a Business Process Management (BPM) System. This poses the challenge of detecting process activities from lowlevel sensor data by means of event abstraction. Existing methods rely on fully supervised approaches, combining sensor data and process event logs to train classification models. However, frequently these logs are not available, invalidating the applicability of such methods. In this paper, we propose a semi-automated approach to detecting process activities from sensor data based on frequently repeated subsequences and automata. We evaluate the approach with a proof-of-concept implementation and data from a small-scale smart factory. Our evaluation demonstrates the applicability of the proposed approach and its effectiveness in detecting process activities from sensor data to support domain experts with data analysis through automated suggestions.Type:conference contributionJournal:Procedia Computer ScienceScopus© Citations 2 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Toward IoT-based Process Analytics: Extending Event Knowledge Graphs with AmbiguityDigital traces of processes executed in the physical world without Information Systems (IS) support are composed of high-level events derived from low-level physical world events (e.g., from IoT). Due to this, frequently these high-level events are ambiguous, i.e., they yield multiple interpretations. As current IS lack ambiguity-awareness, ambiguity in digital traces can compromise process analytics. Motivated by this and current trends toward multi-dimensional analytics, we introduce an ambiguity-aware object-centric representation of digital traces by extending Event Knowledge Graphs with ambiguity. We integrate its construction into a framework to enable analytics for ambiguous IoT-based digital traces. A prototype implementation shows the applicability of the construction from ambiguous process event streams in online settings.Type:conference contributionJournal:Enterprise, Business-Process and Information Systems ModelingScopus© Citations 4 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, A Process to Non-invasively Augment Legacy IoT Systems Using Business Processes and MicroservicesNowadays, Internet of Things (IoT) systems can be found everywhere. Thereby, the IoT systems' software greatly varies in terms of openness and extensibility. Especially older legacy IoT systems only provide limited means of interacting and extending their functionality due to proprietary and closed software interfaces. Nevertheless, IoT systems evolve constantly based on new functional requirements and the need for integration with new systems and IoT devices. In this work we investigate the use of Business Process Management (BPM) technologies in combination with Microservices to augment IoT systems in a non-invasive, lightweight manner focusing on integration, data augmentations, and human-in-the-loop. We start by analyzing the interfaces and processes of a smart factory as a representative legacy system to identify missing functionality, which we generalize into functional requirements for IoT systems. We show how these requirements can be addressed using BPM and microservices based on a discussion of suitable BPM concepts and software architecture, which is accompanied by a prototypical implementation. These results inform our proposal and discussion of a generic development process to non-invasively augment processoriented legacy IoT systems using BPM and microservices.Type:conference paperJournal:International Conference on the Internet of ThingsScopus© Citations 6 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Data-driven Generation of Services for IoT-based Online Activity Detection(2023-11); ; Business process management (BPM) technologies are increasingly adopted in the Internet of Things (IoT) to analyze processes executed in the physical world. Process mining is a mature discipline for analyzing business process executions from digital traces recorded by information systems. In typical IoT environments there is no central information system available to create homogeneous execution traces. Instead, many distributed devices including sensors and actuators produce low-level IoT data related to their operations, interactions and surroundings. We leverage this data to monitor the execution of activities and to create events suitable for process mining. We propose a framework to generate activity detection services from IoT data and a software architecture to execute these services. Our proof-of-concept implementation is based on an extensible complex event processing platform enabling the online detection of activities from IoT data. We use a running example from smart manufacturing to showcase the framework.Type:conference contributionJournal:21st International Conference on Service-Oriented Computing (ICSOC)Scopus© Citations 12 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, An Event-Centric Metamodel for IoT-Driven Process Monitoring and Conformance Checking(2023-09); ; Process monitoring and conformance checking analyze process events describing process executions. However, such events are not always available or in a form suitable for these analysis tasks, for example for manual processes and (semi-)automated processes whose executions are not controlled by a Process-Aware Information System. To bridge this gap, we propose to leverage Internet of Things (IoT) technologies for sensing low-level events and abstracting them into high-level process events to enable process monitoring and conformance checking. We propose an event-centric metamodel for monitoring and conformance checking systems that is agnostic with respect to process characteristics such as level of automation, system support, and modeling paradigm. We demonstrate the applicability of the metamodel by instantiating it for processes represented by different modeling paradigms.Type:conference contributionJournal:21st International Conference on Business Process Management (BPM) WorkshopsScopus© Citations 7 - Some of the metrics are blocked by yourconsent settings
Item type:Publication, Leveraging Digital Trace Data to Investigate and Support Human-Centered Work Processes(Springer, 2024); ; ; ; The ongoing digitization of processes in all domains of everyday life driven by IT systems shows great potential for process automation, analysis, and optimization. In the last decade process mining has advanced to an important and mature discipline of computer science research and has been widely adopted in industry. More recently,-acknowledging the huge potential of digital trace data to study processes-process science has been introduced as an interdisciplinary field studying how processes unfold over time. This paper discusses the potential that arises when using digital trace data not only in the context of highly automated processes but also to investigate humancentered (work) processes and elaborates on associated challenges. Examples range from the semi-automated storage and production processes in a smart factory to healthcare processes to process analysts performing process mining tasks and software engineers reading software artifacts like source code and process models.Type:conference paperVolume:2028Scopus© Citations 5