Improving network security of vehicle-road collaborative system by using artificial intelligence algorithms and bidirectional long short-term memory network

Vehicle cybersecurity is crucial. To address cyberattacks, such as distributed denial-of-service and denial-of-service attacks, faced by vehicle-infrastructure cooperative systems and enhance their security, this study proposes an improved network intrusion detection system that integrates artificial intelligence algorithms with a bidirectional long short-term memory network. Its primary goal is to improve the detection accuracy and robustness of VISs against complex and dynamic cyberattacks. First, a sequential convolutional neural network is used to extract the spatial structural features of network traffic. Second, a bidirectional long short-term memory network is employed to capture the temporal dependence of attack behaviors. Finally, an innovative and improved multivariate gradient optimization algorithm is introduced to dynamically optimize the parameters of sequential convolutional neural network and bidirectional long short-term memory network models during feature extraction and classification, thereby achieving a deep fusion of feature extraction and learning. Compared to existing methods, the sequential convolutional neural network–bidirectional long short-term memory–improved multivariate gradient-based optimization model improves feature representation and model generalization through the improved multivariate gradient-based optimization mechanism. Experimental results demonstrate that this method outperforms mainstream comparison models in key performance metrics such as detection accuracy and F1 score, effectively reducing both false positive and false negative rates, and provides a more efficient and reliable network security solution for vehicle-to-everything cooperative systems. This research demonstrates the significant potential of artificial intelligence algorithms and bidirectional long short-term memory networks to improve the performance of network intrusion detection systems in vehicle-to-everything environments.