Hybrid Blockchain and Deep Learning Model for Robust Internet of Things Security in Intelligent Transportation Systems

Abstract

ABSTRACT Smart cities are digitally advanced urban environments that are equipped with sensor networks to gather, share, and analyze extensive data across interconnected systems. Among various smart city applications, the intelligent transportation system represents one of the most critical and security‐sensitive domains. An intelligent transportation system relies heavily on continuous vehicular communication, a low‐latency decision‐making process, as well as real‐time traffic monitoring. Existing Internet of Things security methods encounter significant computational overhead and limited scalability, making them unfit for real‐time applications. To address these issues, this paper proposes a novel security model, named Deep Residual Stacked Bidirectional Network. The proposed system is integrated into a blockchain‐supported hybrid system to ensure security and privacy for users and systems in smart cities. This enhanced Deep‐Learning model combines the residual learning power with bidirectional long short‐term memory layers. To effectively manage deeper networks, residual connections help mitigate the vanishing gradient problem, while bidirectional long short‐term memory provides sequential dependencies in backward and forward directions. This allows the model to detect patterns in data, especially in security environments where data is highly dynamic and time‐sensitive. Four Internet of Things‐related datasets are used to evaluate the efficiency of the developed algorithm. These datasets offer various real‐world network traffic and attack scenarios that allow comprehensive performance evaluation of the proposed approach in comparison with existing methods. The test outcomes revealed that the blockchain‐supported proposed method outperforms traditional methods with an accuracy of 98.21%, specificity of 97.39%, and F1‐score of 97.46%.