Privacy-Oriented Architecture for Building Automatic Voice Interaction Systems in Smart Environments in Disaster Recovery Scenarios

• Main Authors: Gentile, Antonio Francesco, Macri, Davide, Greco, Emilio, Forestiero, Agostino
• Format: Conference Proceeding
• Language: English
• Subjects: Architecture; Buildings; Disaster Recovery; Machine learning; Neural networks; Privacy; Privacy-preserving; Rasp-berry Pi; Service Continuity; Speech engine; System performance; Training
• ISSN: 2643-6868
• DOI: 10.1109/ICT-DM58371.2023.10286949

In the wake of increasing natural disasters and emergency situations, the deployment of Smart Environments with automatic voice interaction systems has become vital for efficient disaster recovery efforts. However, the widespread adoption of such systems raises significant concerns about user privacy and data security. This paper proposes a novel privacy-oriented architecture tailored explicitly for building automatic voice interaction systems in Smart Environments, designed to address the privacy challenges inherent in disaster recovery scenarios. The proposed architecture emphasizes the implementation of robust privacy-preserving mechanisms, ensuring that sensitive user data remains secure and inaccessible to unauthorized entities. To demonstrate the practicality and effectiveness of the proposed architecture, a proto-type voice interaction system was developed and deployed in a simulated environment. The system's performance is evaluated against traditional architectures, showcasing its superior ability to safeguard user privacy while delivering accurate and responsive voice-based services.