Blockchain and deep learning-enabled IoT device-to-device authentication approach for smart cities using 5th generation technology

【Author】 Chandran, K. Prabhu; Bhuvaneswari, P.; Sivasankaran, V.; Vimala, S.

【Source】ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE

【影响因子】7.802

【Abstract】As smart cities evolve, the increasing number of Internet of Things (IoT) devices requires secure authentication mechanisms for device-to-device (D2D) communication, especially within 5G networks. To address security challenges in D2D communication, Blockchain (BC) technology is utilized. Existing methods face issues like single-point failure attacks, high implementation costs, and data privacy concerns. This work aims to develop a secure and efficient authentication model leveraging BC technology and a Deep Q Network (DQN) for key generation to enhance IoT security in smart city applications. A novel authentication model, Deep Q Network-SecAuth (DQN-SecAuth), is proposed for secure D2D communication using BC. The model involves key entities such as the Registration Authority Center (RAC), blockchain, and IoT devices. The authentication process includes six stages: setup, registration, key generation, authentication, new device addition, and formal verification. In the key generation stage, the secret key is generated using a Deep Learning (DL) model named DQN. The proposed model also incorporates encryption, hash functions, Chebyshev polynomials, and XOR operations to ensure security. The DQN-SecAuth model achieved a minimum computational time of 9.525 s and memory usage of 42.897 megabytes (Mb), consensus delay of 0.420 s, energy consumption of 0.715 J, latency of 0.469 s, power consumption of 0.270 kW-hour (kWh), and throughput of 0.804 megabits per second (Mbps). The key novelty of this work is the use of a DQN for adaptive key generation, integrated into a BC-based authentication framework. This combination enhances security, reduces computational overhead, and outperforms traditional static key-generation and authentication methods in smart city IoT environments.

【Keywords】Data encryption; Key generation; Internet of things; Blockchain; Deep Q network

【发表时间】2025 DEC 9

【收录时间】2025-09-20

【文献类型】

【主题类别】

--