Node Fault Prediction Assisted Small-World IoT Networks Using ML Frameworks: Towards Performance Improvement

The rapid growth of the Internet of Things (IoT) networks has led to the deployment of large-scale networks, enabling seamless connectivity and data exchange among various devices. To manage the complexity and ensure efficient communication in these expansive networks, adopting suitable network architecture becomes crucial. Small-world networks, characterized by a high Average Clustering Coefficient (ACC) and low Average Path Length (APL), have emerged as a promising architecture for IoT systems due to their efficient communication. However, introducing the Small-World Characteristics (SWC) in an IoT network is challenging due to the need for strategic placement of long-range connections while maintaining low APL and high ACC and ensuring scalability and robustness. In this work, we introduce SWC into the network using an actor-critic reinforcement learning algorithm. Additionally, ensuring the reliability of sensor nodes is crucial to maintaining the overall network performance. Therefore, we propose a joint method for dynamic node fault prediction and data routing within small-world IoT networks using advanced Machine Learning (ML) frameworks. Several data routing experiments have been conducted to validate the effectiveness of the proposed approach using simulated small-world IoT networks. We analyzed major network parameters such as lifetime, latency, and throughput. We compared the proposed method with existing state-of-the-art approaches and observed promising results.