Block-HAR: Trusted Human Activity Recognition Framework for 5G-assisted Smart Healthcare

The paper introduces the Block-HAR framework, that integrates blockchain and cloud computing at the backdrop of Fifth Generation (5G) services to ensure seamless interactivity among healthcare providers. This framework consists of two layers: the data collection layer, and the blockchain and Smart Contracts (SC) layer. At data collection layer, sensors are placed at key points to monitor physiological changes or movement patterns of aged and differently-abled patients, triggering alerts for deviations. The collected data is transmitted via the 5G Ultra Reliable Low Latency Communications (uRLLC) service to the Google cloud for storage and analysis. An ensemble model is proposed to process sensor data within the cloud, identifying critical movement or medical patterns. At the blockchain and SC layer, activity reports are securely shared with stakeholders via smart contracts, while trained data are stored offline in the Interplanetary File System (IPFS), accessible through a content key. The proposed framework achieves an average accuracy of 98.13%, a latency improvement of 94.12% over Wi-Fi 6, and a Block Error Rate (BLER) improvement of 92.17% compared to Wi-Fi 6, demonstrating significant enhancements in real-time healthcare monitoring and intervention.