Analysis of DoS Attacks and Detection Techniques in Smart City Systems

In smart city systems, Denial of Service (DoS) attacks threaten critical infrastructure reliability. This paper analyzes DoS attack simulations and evaluates four detection models: Support Vector Machine (SVM), Random Forest (RF), Convolutional Neural Network (CNN), and Recurrent Neural Network (RNN-LSTM). RF emerged as the most effective model, achieving over 99% accuracy, precision, recall, and F1-scores across all attack intensities, making it ideal for real-time scenarios due to its efficiency and balanced resource usage. While SVM performed well in low to medium-intensity attacks, its high CPU usage limits real-time applicability. CNN achieved 94 % accuracy in spatial anomaly detection but exhibited reduced recall under high-intensity attacks. The RNN-LSTM excelled in high-intensity scenarios, recording the highest recall (97<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>), but its extended training time and memory usage pose challenges. These findings underscore the need for hybrid approaches like CNN-LSTM to optimize accuracy and efficiency in smart city cybersecurity.