Edge-Based IoT Security Using Lightweight Cryptography and Machine Learning for Smart City Communication Systems

Abstract

Smart city communication systems increasingly depend on Internet of Things (IoT) devices for real-time monitoring, automation, traffic management, smart healthcare, intelligent surveillance, and energy optimisation. However, IoT networks remain highly vulnerable due to limited computing resources, large-scale device heterogeneity, insecure communication channels, and the growing sophistication of cyberattacks. Traditional cloud-centric security frameworks often introduce high latency, bandwidth overhead, and increased exposure to threats. To overcome these challenges, edge-based security solutions are gaining prominence by enabling real-time protection closer to the data source. This research paper proposes an integrated edge-based IoT security model that combines lightweight cryptography for secure communication and machine learning (ML) for anomaly and intrusion detection in smart city environments. The system design is evaluated using a simulated smart city IoT network with varied attack scenarios including distributed denial of service (DDoS), spoofing, botnet infiltration, and data manipulation. Statistical analysis demonstrates that the integrated model significantly improves detection performance while maintaining low computational overhead, making it suitable for resource-constrained devices. The results indicate strong improvement in accuracy, reduced latency, and enhanced resilience against major IoT threats. This work provides a scalable and efficient security framework for next-generation smart city communication systems by merging cryptographic integrity and intelligent edge monitoring.

Introduction

The text presents a comprehensive study on securing smart city IoT communication systems using an edge-based security architecture that combines lightweight cryptography and machine learning (ML). Smart cities depend on interconnected IoT devices for transportation, healthcare, energy, and public services, generating massive real-time data streams that demand secure, low-latency communication. However, IoT devices are highly vulnerable to cyberattacks due to limited resources, weak authentication, and inadequate updates. Traditional cloud-based security solutions introduce latency, bandwidth overhead, and privacy risks, making them unsuitable for time-critical smart city applications.

To address these issues, the paper proposes an edge-centric security framework with three layers:

1. IoT Device Layer using lightweight encryption for constrained devices,

2. Edge Security Layer that performs local authentication, encryption, intrusion detection, and real-time mitigation using ML, and

3. Cloud Layer for large-scale analytics, storage, and model training.

The architecture employs lightweight cryptographic techniques (such as Ascon-based AEAD, hash-based authentication, and symmetric session keys) to ensure confidentiality, integrity, and authentication with minimal computational overhead. In parallel, edge-based ML intrusion detection systems analyze traffic features like packet rate, payload size, and authentication failures to detect attacks such as DDoS, spoofing, and botnets in near real time.

Simulation-based evaluation shows that Random Forest models achieve the highest detection accuracy (97.4%) with low false positives. Edge-based detection significantly reduces latency (58 ms) compared to cloud-only IDS (210 ms), making it suitable for safety-critical smart city functions. Lightweight cryptography also demonstrates much lower processing time than traditional encryption methods.

Overall, the study concludes that integrating lightweight cryptography and ML-based intrusion detection at the edge provides a scalable, efficient, and low-latency security solution for smart city IoT networks. While challenges such as model drift, training data quality, and key management remain, the proposed layered approach significantly enhances resilience against both external and insider threats in smart city environments.

Conclusion

This research paper proposed an edge-based IoT security framework for smart city communication systems by integrating lightweight cryptography and machine learning intrusion detection. The results demonstrate that lightweight AEAD encryption ensures secure device-edge communication with minimal overhead, making it suitable for constrained IoT nodes. In parallel, edge-based ML monitoring improves real-time detection of DDoS, spoofing, botnet, and injection attacks while reducing latency compared to cloud-only security approaches. Statistical analysis confirms strong detection performance, especially with Random Forest models, and highlights the importance of low false positive rates for smart city reliability. The proposed solution offers a scalable, low-latency, and energy-efficient security architecture suitable for future smart city deployments. Future work may focus on adversarial robustness, federated model training, dynamic key distribution, and trust-based edge collaboration mechanisms.

References

[1] Alrawais, Alaa, et al. “Fog Computing for the Internet of Things: Security and Privacy Issues.” IEEE Internet Computing, vol. 21, no. 2, 2017, pp. 34–42. [2] Bertoni, Guido, et al. “The Sponge Functions Corner.” Cryptographic Hardware and Embedded Systems (CHES), Springer, 2007. [3] Ferrag, Mohamed Amine, et al. “Security for 5G and IoT Networks: A Survey.” Computer Networks, vol. 183, 2020. [4] Kolias, Constantinos, et al. “DDoS in the IoT: Mirai and Other Botnets.” Computer, vol. 50, no. 7, 2017, pp. 80–84. [5] NIST. “Lightweight Cryptography Project.” National Institute of Standards and Technology, 2023. Perrig, Adrian, et al. “SPINS: Security Protocols for Sensor Networks.” Wireless Networks, vol. 8, 2002, pp. 521–534. [6] Powers, David M. W. “Evaluation: From Precision, Recall and F-Measure to ROC.” Journal of Machine Learning Technologies, 2011. [7] Roman, Rodrigo, Javier Lopez, and Masahiro Najera. “Securing the Internet of Things.” Computer, vol. 44, no. 9, 2011, pp. 51–58. [8] Satyanarayanan, Mahadev. “The Emergence of Edge Computing.” Computer, vol. 50, no. 1, 2017, pp. 30–39. [9] Shi, Weisong, et al. “Edge Computing: Vision and Challenges.” IEEE Internet of Things Journal, vol. 3, no. 5, 2016, pp. 637–646. [10] Sicari, Sabrina, et al. “Security, Privacy and Trust in IoT: The Road Ahead.” Computer Networks, vol. 76, 2015, pp. 146–164. [11] Zanella, Andrea, et al. “Internet of Things for Smart Cities.” IEEE Internet of Things Journal, vol. 1, no. 1, 2014, pp. 22–32. [12] Conti, Mauro, et al. “Internet of Things Security and Forensics.” Future Generation Computer Systems, vol. 78, 2018, pp. 544–546. [13] Li, Shancang, et al. “Secure and Energy-Efficient Transmission for IoT.” IEEE Transactions on Industrial Informatics, vol. 14, 2018. [14] Kumar, Neeraj, and Jong-Hyouk Lee. “Blockchain and IoT Security.” IEEE Communications Surveys & Tutorials, vol. 22, 2020. [15] Nguyen, Thanh, et al. “Deep Learning for IoT Intrusion Detection.” Future Internet, vol. 12, no. 10, 2020. [16] Moustafa, Nour, and Jill Slay. “UNSW-NB15 Dataset.” Military Communications and Information Systems Conference, 2015. [17] Doshi, Rohan, Noah Apthorpe, and Nick Feamster. “Machine Learning DDoS Detection for IoT.” IEEE Security & Privacy Workshops, 2018. [18] Hossain, Md. Shamim, et al. “Securing Smart Cities Using AI.” IEEE Access, vol. 7, 2019. [19] Rahman, Md. Arifur, et al. “Lightweight Authentication Protocols for IoT.” Sensors, vol. 20, 2020. [20] Ali, Zafar, et al. “Edge-Based Security for Smart Cities.” IEEE Access, vol. 9, 2021. [21] Xu, Xiang, et al. “Edge Intelligence for IoT Security.” IEEE Network, vol. 34, 2020.

Copyright

Copyright © 2026 Dr. Ennam Govinda, Dr. C.P.N.J Mohan Rao, Nattala Nagamani, P. Raghava Kumari. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.