Detecting Blackhole Nodes in IoT Environment using Logistic Regression-Based Trust-Based Security (LRTS)

Black hole attacks pose a critical threat to Internet of Things (IoT) networks by maliciously dropping data packets, severely disrupting communication, and compromising network reliability. These attacks exploit the lack of robust security mechanisms in common routing protocols, leading to significant data loss and network performance degradation. Addressing this vulnerability is essential to ensure the secure and efficient operation of IoT systems. This paper proposes an intelligent detection mechanism leveraging logistic regression (LR) and trust metrics to predict whether a node exhibits black hole behavior. Trust parameters such as end-to-end delay (ED), packet delivery ratio (PDR), energy consumption (EC), and reputation trust (RT) are employed as features in a LR model. The model is trained using stochastic gradient descent (SGD), initializing coefficients and a threshold for integrated trust (IT), and iteratively updating them based on the dataset. In the prediction phase, the learned coefficients compute IT values for unseen nodes, which are then classified as Trusted or Malicious. Simulation results confirm the effectiveness of the proposed model, achieving 85% accuracy—demonstrating a 13% improvement over conventional models with 72% accuracy. This proactive and data-driven approach significantly enhances IoT network resilience by enabling early detection and mitigation of black hole attacks.