A blockchain-based secure data transmission framework in IoT using adaptive deep network with optimized cryptography...

This work presents a secure data transmission process in the Internet of Things (IoT). Initially, the required data are collected and given to the Adaptive and Sparse Attention-based Dense Long Short-Term Memory (ASA-DLSTM) network for intrusion detection. The adaptive nature of the model allows for optimizing the parameters using the Sorted Fitness-based Addax Optimization Algorithm (SF-AOA). Once intrusions are detected, the data is used for the data transmission phase. It is performed using Optimal Key-based Elliptic Galois Cryptography (OK-EGC). By combining Elliptic with Galois fields and an optimal key management strategy, the proposed OK-EGC method enhances both encryption efficiency and security. Moreover, the integration of optimal key-based management using the same SF-AOA ensures that cryptographic keys are dynamically optimized based on the network's security requirements. Then, the effectiveness of the model is compared with existing systems. The accuracy of the implemented SF-AOA-ASA-DLSTM technique is 95.97%, which is higher than the conventional techniques, such as DNN (83.77%), SVM (83.19%), 1DCNN (90.26%), and ASA-DLSTM (93.6%) for the batch size value 64. Thus, the results display that the designed model addresses the critical challenges of IoT data security by providing both robust intrusion detection and secure communication.