Theoretical results:
- We design an expandable simulation model with real-data validation to generate justifiable datasets.
- We propose a machine learning-based approach to detect jamming attacks, specifically on BSMs in vehicular networks. The detection is made based on medium access observation at the individual vehicle. The approach is evaluated and analyzed in distinguished training datasets. 
Training results: 01 Master thesis
+ Master student: Huỳnh Vinh Nam - USTH
+ Thesis title: “Automatic landmarks detection on Insect wings using Deep Learning method”
+ Supervisor: TS. Nguyễn Hoàng Hà (project member)

In vehicular networks, information included in safety messages is essential and urgently needed for all vehicles to react for their safety. It requires vehicles to broadcast these short messages periodically. Therefore, they are prone to be a target for simple radio attacks such as jamming. The degradation in the quality of communication under jamming attacks is similar to lousy condition communication. Therefore, detection is challenging, especially when many vehicles join the network. The most recent detection method based on medium access observations can detect most attacks. However, false alarms are also frequently raised due to the wrong estimation of the number of neighbors. In this work, we propose to apply machine-learning techniques to exploit the medium access information collected by individual vehicles to improve the previous methods. The machine learning-based approach requires a relevant dataset which is also a challenge in a vehicular network. 
The simulation model and datasets should enable further research on jamming attacks as they provide relevant data in vehicular networks. We believe these models, datasets, and their use in a machine learning-based approach will interest researchers who work on the topic of security on vehicular networks but encounter the challenges of collecting data.