Semantic Low-Code Data Fusion on the Edge to Support Intelligent Traffic Management

Poster presented at the Transport Research Arena (TRA) 2026 for the paper 'Semantic Low-Code Data Fusion on the Edge to Support Intelligent Traffic Management'

Abstract: Intelligent traffic management applications require integrating data from diverse sources on the road infrastructure like radar sensors, cameras, and connected vehicles. The growing computational power deployed on the edge enables distributed computation of the data and near-real-time traffic control. However, interoperability is hindered by two key challenges: heterogeneous protocols for data access and inconsistent data semantics across sources. A data harmonisation solution is needed to retrieve data from different data sources while enabling a uniform representation of the information for data fusion and integrated analysis. The paper describes our solution to support the low-code definition of data integration pipelines by configuring a set of modular and reusable components. We leverage Semantic Web technologies and declarative mapping rules to resolve semantic inconsistency through a common conceptual model. A set of deployment templates enables the execution of the pipelines both on the cloud or on the edge to support diverse downstream applications with different performance and scalability requirements. Finally, the paper discusses the testing and validation of such a solution to support the reduction of collisions caused by unsafe braking at road intersections in a live deployment at the Mobility Lab in Helsinki.