Design of mTCN framework for disaster prediction a fusion of massive machine type communications and temporal convolutional networks

Abstract

Natural disasters such as floods, tsunamis, and earthquakes significantly impact lives and infrastructure, highlighting the urgent need for accurate and real-time prediction systems. Current methods often suffer from limitations in scalability, privacy, and real-time data integration, particularly in large-scale disaster scenarios. This study introduces the mTCN-FChain framework, a novel solution that combines Massive Machine-Type Communications (mMTC) and Temporal Convolutional Networks (TCNs) with federated learning and blockchain technology. The objective is to develop a scalable, secure, and efficient system for real-time disaster prediction using IoT data streams. Lightweight edge-based TCNs enable localized anomaly detection, while federated learning ensures privacy-preserving collaborative model training across edge devices. Blockchain integration secures model updates and provides traceability. Using datasets for earthquakes, floods, and tsunamis, the framework was implemented in Python and validated using metrics such as MAE, MSE, and RMSE. The results show significant performance improvements over existing methods, with higher accuracy, reduced latency, and robust scalability in disaster prediction tasks. Tools like TensorFlow, PyTorch, and Hyperledger Fabric were employed for implementation. The study concludes that mTCN-FChain offers a transformative approach to disaster resilience, though future work will focus on optimizing blockchain integration and expanding real-world applicability to enhance its robustness and adaptability.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.