Systems architecture for scalable multi-agent building control: A semantic ontology and fog computing framework

Abstract

Building automation systems typically employ numerous distributed Direct Digital Controllers (DDCs) requiring individual programming and maintenance, creating deployment complexity that limits advanced control adoption. While centralized fog-level computing approaches offer potential benefits, they require validation of real-time performance capabilities. This work investigates the feasibility of centralized fog-level control for building automation through a systematic architecture addressing semantic data processing, performance requirements, and automated configuration. This is the first step towards replacing DDCs. We present a control-oriented semantic ontology with automated contextualization capabilities, achieving 122-138 ms response times suitable for room-level building automation applications through fog-level serverless computing. The framework demonstrates automated configuration for occupant-facing systems in an office building, providing evidence for the viability of centralized approaches in building control applications.