Metaverse Driven Edge-fogging-cloud Network for Complex Human Activity Recognition using Sensors Fusion

Metaverse and digital twins are promising for applications in people-centered domains, such as smart cities and telemedicine. Meanwhile, combining the technologies with human activity recognition (HAR) can substantially improve identification accuracy and computational efficiency. Besides, using an edge-fog-cloud network architecture for data transmission offers advantages regarding storage space and computational resources on edge devices, thereby enhancing analysis efficiency. We present a novel HAR system that leverages the edge-fog-cloud network architecture, which achieves three key functionalities. First, the system integrates data from multimodal sensors to address the computational challenges and data transmission latency associated with complex HAR tasks. Second, the Kafka network architecture and the concept of digital twins ensure secure and timely transmission within the system. Third, the system employs hierarchical algorithm models to enhance HAR prediction accuracy while reducing computational time at each layer.