A multi-agent architecture for context sources integration in smart cities

Abstract

Contextual data in smart cities are present in large quantities and distributed sources. Many applications can benefit from these data to provide better services to their users. The scale and dynamic nature of urban environments pose significant challenges in making context sources available to applications. These challenges involve transparent access to context, resilience, decentralization, extensibility, scalability, and redundancy of data. This study introduces a new architecture designed to address these issues. This architecture aims to facilitate the acquisition of context by integrating distributed data sources. The developed architecture not only overcomes the challenges posed by the scale and dynamicity of urban environments but also prepares for more innovative and effective solutions for smart cities. The architecture is distributed, decentralized, and fault-tolerant, providing data fusion mechanisms and dynamic context source composition. Compared to existing works, our architecture contributes to the state-of-the-art addressing all these five challenges in one design. The architecture uses the multi-agent paradigm, which is inherently distributed and facilitates decentralization. A scenario was used to execute several experiments demonstrating that the architecture can obtain context data transparently by any application.