An Optimal Storage Cyber Security for Block-Chain-Assisted WSN: Identifying Malicious Nodes and Enhancing the Storage of Data

Wireless sensor networks (WSN), an crucial part of the Internet of Things (IoT), are utilized in cybersecurity, healthcare, and military applications but face challenges like resource limitations and rogue nodes, which affect security. This study proposes a WSN architecture with three types of nodes: base stations (BSs), cluster heads (CHs), and sensor nodes (SNs). SNs collect data, which CHs process and send to BSs. Both SNs and CHs register using blockchain with the Practical Byzantine Fault Tolerance (PBFT) consensus for authentication and data security. After registration, nodes transmit data securely to BSs. The WSN dataset (WSN-DS) with min-max normalization is used, and a new method, ε-insensitive two-plane support vector regression (ε-TSVR), identifies malicious nodes. The findings show the proposed method improves accuracy (91%), f1 score (93%), recall (94%), and precision (92%), offering a more secure WSN solution for IoT environments, even with malicious nodes.