New deep learning system CANGuard uses CNN-GRU-attention hybrid to detect cyberattacks on vehicle CAN bus networks

CANGuard combines Convolutional Neural Networks (CNN), Gated Recurrent Units (GRU), and attention mechanisms to identify intrusion attacks in connected vehicles

System targets critical security threats including Denial-of-Service (DoS) and spoofing attacks on Controller Area Network (CAN) buses that could disable vehicle communication

Model trained and evaluated on CICIoV2024 dataset to protect Internet of Vehicles (IoV) infrastructure in smart transportation systems

Spatio-temporal architecture designed to enhance vehicle safety by preventing system malfunctions and loss of control from cyberattacks