Publikation
The Challenges of Multimodal IoT Data Fusion for Process Understanding and Monitoring
Zahra Ahmadi; Sönke Knoch; Massimiliano Sampaolo; Mohsen Shirali
In: Andrea Maldonado; Yannis Bertrand; Ronny Seiger (Hrsg.). Business Processes Meet the Internet of Things 2026. International Workshop on Business Processes Meet the Internet-of-Things (BP-Meet-IoT-2026), located at EDOC 2026, September 15, Enschede, The Netherlands, CEUR Workshop Proceedings, 11/2026.
Zusammenfassung
Multimodal IoT data fusion offers a promising way to improve process understanding and monitoring by combining heterogeneous IoT data across sensor, visual, acoustic, spatial, and contextual modalities. However, the integration of heterogeneous sources introduces several challenges, such as differences in data formats, sampling rates, and levels of abstraction. Despite the growing availability of multimodal IoT infrastructures, a unified framework that
systematically addresses these challenges remains largely absent in the literature. Therefore, this paper focuses on the key challenges of multimodal IoT sources for process understanding and monitoring. It identifies four key sub-challenges: heterogeneous modality integration, fusion strategy selection, real-time activity and anomaly inference, and the construction of holistic and explainable process views. These sub-challenges are examined by reviewing research across various domains. Based on this analysis, the paper identifies key research gaps and opportunities and proposes a forward-looking approach for developing reliable and explainable multimodal IoT process intelligence systems.