Prototype of a multi-agent architecture for real-time monitoring of personal protective equipment using an RFID-based smart gate

Ensuring operator health and safety in complex industrial environments requires continuous verification of Personal Protective Equipment (PPE) usage. This work presents a prototype of a distributed monitoring framework based on the MADS-NET multi-agent architecture, integrating heterogeneous RFID sensors for real-time PPE detection and authentication at a smart gate. The gate implemented using a Raspberry Pi 5, hosts dedicated agents that collect data from High-Frequency (MFRC522) and Ultra-High-Frequency (UHF) RFID readers, operating at different sampling rates and detection ranges. These agents locally assess PPE compliance against predefined thresholds and communicate via a wireless broker with temporal synchronization, enabling a scalable and modular design that can be extended to multiple gates. A dedicated node handles real-time visualization of detection events and signal strength metrics, while a logger node stores data in a no-SQL database for auditing and historical analysis with configurable logging frequency. The prototype demonstrates low latency in data exchange, high repeatability in detection, and high adaptability to different industrial layouts. Performance indicators (including response time, read rate, and false positive/negative rates) confirm the measurement reliability of the system. Preliminary results validate the feasibility of the proposed decentralized approach with quantitative accuracy metrics, paving the way for future deployment of a fully distributed, multi-gate safety monitoring system in intelligent manufacturing environments.

Keywords: Multi-agent system, Personal Protective Equipment (PPE), Real-time monitoring, Industrial safety, Heterogeneous RFID sensors.