Semantics for Enhancing Communications- and Edge-Intelligence-enabled Smart Sensors: A Practical Use Case in Federated Automotive Diagnostics

Modern edge artificial intelligence (AI) chipsets and edge-intelligence-enabled smart sensors frameworks support real-time data processing and event detection at the signal source. Beyond measuring local conditions and transmitting corresponding signals, AI-enabled smart sensors provide the capability to interpret and analyze signals through local analytical operations. Its well-known semantics provide an abstraction layer for better comprehension and self-descriptive applications. This paper identifies and explains the advantages of semantics and ontology engineering in integrating edge AI-enabled smart sensors and their applications into federated automotive systems. Building on demonstrated work using AudioSet for event detection, the application presented as an AI mechanic at the edge demonstrates how smart sensor integration necessitates the re-evaluation of how automotive diagnostic trouble codes (DTCs) are generated and processed. While DTCs and their logic are intended to facilitate correct assembly and repair, they do not always clearly indicate the underlying issue. This paper also underlines the need to review ontologies and semantic models for smart sensor standard specifications in light of AI-enabled smart sensors.