Federated Learning for Smart City Network Attack Detection and Classification: A Literature Review

The growth of smart cities and the Internet of Things (IoT) has generated massive volumes of sensitive data, creating significant cybersecurity challenges and limitations for centralized machine learning models. Federated Learning (FL) has emerged as a decentralized paradigm that preserves privacy by training models locally on devices. This systematic literature review aims to analyze current trends, challenges, and solutions in applying FL for cybersecurity in smart city environments based on recent publications. The analysis reveals that FL can be effectively implemented in Intrusion Detection Systems (IDS) to detect various types of network attacks. A major universal challenge is the presence of Non-Independent and Identically Distributed (Non-IID) data, which can degrade model performance. The key findings highlight that the most effective implementations of FL are not standalone, but rather hybrid approaches that integrate FL with complementary technologies—such as blockchain to enhance data integrity, adaptive algorithms to handle Non-IID data, and additional privacy-preserving techniques like Differential Privacy and encryption. Overall, the development of frameworks that combine FL with supporting technologies shows strong potential for enhancing the security of smart cities.