Blockchain-Enabled Security Framework for Malicious Node Detection and DDoS Defense in Agricultural IoT

With the rapid development of internet of things (IoT) technologies in the consumer-electronics domain, intelligent sensor networks face increasingly severe security challenges. To address threats such as selective-forwarding, data tampering, and distributed denial-of-service (DDoS) attacks in agricultural IoT environments, this paper proposes a trustworthy sensor-data storage model based on a four-layer data perception-processing-storage-application architecture. At the data perception layer, blockchain technology is integrated to ensure data traceability. At the data processing layer, we design a cluster-head election mechanism that combines K-means clustering, a trust evaluation mechanism, and energy-consumption optimization, yielding an 85.7% extension in network lifetime compared with the LEACH protocol. We further propose a blockchain-based malicious node detection method (BMND) that maintains 94.6% detection accuracy even when 30% of nodes are malicious. Targeting agricultural field data characteristics, a joint preprocessing strategy based on TF-IDF and the information-gain ratio was developed, achieving a 91.5% noise-removal rate. In addition, a DDoS defense model that fuses a multimodal deep neural network with the elastic weight consolidation (EWC) algorithm sustains 83.5% recognition accuracy under 80% high-intensity attack conditions. Experimental results show that the proposed security mechanism exhibits excellent adaptability and stability across different sensor-density settings, providing reliable protection for edge-intelligent devices in the consumer-electronics domain and offering theoretical and practical guidance for secure architecture design of intelligent device networks.