Internet of Vehicles internal intrusion detection method based on abnormal behavior discovery

• Main Authors: XING LING, HWANG WON-HO, WANG KUN, ZHAO PENGCHENG, WU HONGHAI
• Format: Patent
• Language: Chinese; English
• Published: 09.05.2023
• Subjects: ELECTRIC COMMUNICATION TECHNIQUE; ELECTRICITY; TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHICCOMMUNICATION

The invention discloses an Internet of Vehicles internal intrusion detection method based on abnormal behavior discovery, and the method comprises the steps: constructing an Internet of Vehicles internal intrusion detection model based on a memory enhanced auto-encoder, and enabling the model to achieve the intrusion detection based on a principle that the reconstruction error of abnormal communication behavior data is larger than that of normal communication behavior data. And the cloud server initializes related parameters of the Internet of Vehicles internal intrusion detection model and issues the related parameters to each roadside unit, the Internet of Vehicles internal intrusion detection model is trained based on federal aggregation, and after training is completed, each roadside unit performs intrusion detection on vehicles in a coverage range by using the trained Internet of Vehicles internal intrusion detection model. According to the method, the communication behavior characteristics of the normal