A novel effective key synchronization approach based on optimized deep neural networks for IoT-based low-power wide area networks

Abstract

Low-Power Wide Area Networks (LPWANs) represent a category of wireless technologies hailed for their efficiency in facilitating communication for Internet of Things (IoT) applications. This efficacy is attributed to their characteristics of low power consumption, extensive wireless transmission range, and cost-effectiveness. Despite these notable advantages, LPWANs exhibit drawbacks such as limited processing power, modest transmission rates, and notably constrained payload sizes, posing challenges for encryption techniques. The inadequacy of existing cipher-chaining encryption methods for LPWANs is underscored by their dependency on high computing power and payload capacity. To address this issue, this paper introduces an innovative chaining encryption approach tailored for LPWAN IoT technology. The proposed method incorporates a key synchronization mechanism based on deep learning algorithms. The effectiveness of this approach has been rigorously assessed through case studies and experiments. The experimental results demonstrate the commendable performance of the proposed approach: i) The proposed method achieves an average effort of 1.0202 for key synchronization after 10 lost packets, compared to 1.0215 for the best existing method; ii) It attains a 98.45% success rate in key synchronization after 10 lost packets, while the best existing method achieves 98.19%; iii) The proposed approach ensures 98.27% accuracy in receiving messages correctly under conditions of completely random data receipt, compared to 98.22% for the best existing method. These results establish the proposed method as a highly competitive solution among encryption approaches for LPWANs.