This paper presents an adaptive software architecture designed for edge computing in crowd-monitoring applications utilizing Wi-Fi fingerprinting. The architecture enables the real-time detection and analysis of mobile devices in proximity, facilitating crowd density estimation and management in various scenarios, such as tourist destinations, event venues, and public spaces. The adaptive features of the architecture include auto-healing, vertical handover, and over-the-air updates. By leveraging self-configuration capabilities, the system dynamically adjusts communication protocols between Wi-Fi and LoRaWAN to ensure resilience and continuity of data transmission under changing network conditions. Furthermore, over-the-air update mechanisms facilitate seamless deployment of software updates to edge nodes, enhancing security and performance over time. By employing a Wi-Fi fingerprinting algorithm, the system accurately differentiates mobile devices while preserving user privacy in compliance with GDPR. Integration with cloud-based servers and APIs enables seamless data ingestion and visualization, empowering destination managers with actionable insights for effective crowd management. Overall, this adaptive edge computing architecture represents a significant advancement in crowd-monitoring technology, offering scalability, flexibility, and privacy for diverse application scenarios.

This work was supported by project 2024.07624.IACDC/2024, funded by FCT – Fundação para a Ciência e a Tecnologia, I.P., under PRR – Plano de Recuperação e Resiliência funding, investment REC05-i08 - Ciência Mais Digital.

Adaptive Software Architecture,Edge Computing,Crowd Monitoring,Wi-Fi Fingerprinting,Auto-Healing,Vertical Handover,Over-the-Air Updates,Privacy

• Computer and Information Sciences - Natural Sciences
• Electrical Engineering, Electronic Engineering, Information Engineering - Engineering and Technology