Clinical Report: Edge-Cloud Collaboration Enhanced by Reinforcement Learning
Background
The Internet of Medical Things (IoMT) represents a significant advancement in healthcare, enabling continuous monitoring and proactive patient care. However, challenges such as data privacy, real-time processing, and energy efficiency remain critical due to the limitations of edge devices.
Data Highlights
The proposed system demonstrated a significant reduction in latency for critical alerts and improved battery life of IoT nodes to 8.5 days compared to conventional methods, as validated by experimental results.
Key Findings
Federated Variational Mode Decomposition (VMD) ensures high-fidelity feature extraction while preserving data privacy.
A SparseBonsai neural network enables real-time classification of medical signals on resource-constrained sensor nodes.
A Proximal Policy Optimization (PPO) reinforcement learning agent optimizes data processing decisions based on severity, network conditions, and battery levels.
The advanced Sha-Dragon optimization algorithm enhances energy efficiency through effective resource and transmission power allocation.
Security is reinforced through a dual-layer approach using ASCON v1.2 and WireGuard VPN with ChaCha20-Poly1305 encryption.
Clinical Implications
The framework offers a solution for real-time healthcare monitoring while addressing issues of data privacy and energy efficiency.
Conclusion
The proposed hierarchical framework addresses the challenges of IoMT by enhancing security, energy efficiency, and real-time processing capabilities.
