Enhanced Malicious Node Detection in Wireless Sensor Networks Using Multimodal Contrastive Domain Sharing Generative Adversarial Networks and Blockchain‐Based Distributed Storage

ABSTRACT Wireless sensor networks (WSNs) suffer from the risk of security breaches because of limited node resources, dynamic topologies, and lack of security mechanisms. These limitations affect the integrity of the data and the reliability of the network. Therefore, this paper proposes an enhanced malicious node detection in wireless sensor networks using multimodal contrastive domain sharing generative adversarial networks and blockchain based distributed storage (MND‐MCDSGAN‐WSN). Firstly, the sensor data gathered from the wireless sensor networks dataset are used. The collected data are cleaned with the Multi‐Observation Fusion Kalman Filter (MOFKF) to remove the noise and to make the data more consistent. The Multimodal Contrastive Domain Sharing Generative Adversarial Networks (MCDSGAN) framework gets preprocessed data as input and classifies the malicious activities as Normal, Grayhole, Blackhole, TDMA, and Flooding. Meanwhile, the Adaptive Marine Predator Optimization Algorithm (AMPOA) selects the MCDSGAN model parameters to improve the stability and the generalization of the model. The Interplanetary File System (IPFS) is used to keep the verified node data safe, and a fair proof‐of‐reputation (FPoR) consensus mechanism is employed to ensure trustful nodes' participation and updates that are resistant to tampering in the blockchain. The combined architecture enhances detection reliability, thus helping trust management and giving a scalable security solution to WSNs that are underresourced. The proposed MND‐MCDSGAN‐WSN method achieves higher accuracy of 99.07% when evaluated with other existing methods.