| KEYWORD |
CAN vehicle Network IDS using Neural Network
keywords AUTHENTICATION, AUTOMOTIVE, CAN NETWORK, IDENTITY MANAGEMENT, INTRUSION DETECTION, NEURAL NETWORKS, SECURITY
Reference persons STEFANO DI CARLO, ALESSANDRO SAVINO
External reference persons OBERT FRANCO
Research Groups DAUIN - GR-24 - SMILIES - reSilient coMputer archItectures and LIfE Sci
Thesis type APPLIED, APPLIED RESEARCH, EXPERIMENTAL RESEARCH, HARDWARE DESIGN
Description Automotive connectivity is no longer an afterthought. With the growth of complex customer features, the automotive Electronic Control Units (ECUs) are increasing their connectivity ability, not only Vehicle-to-Vehicle (V2V) but also Vehicle-to-Everything (V2X). In this condition, the automotive market is becoming the perfect playground for hackers who want to exploit in-vehicle network protocols’ vulnerabilities. The primary automotive communication protocol is the Controller Area Network (CAN). Limited hardware resources and real-time environment lack authentication and integrity mechanisms, making the vehicle network vulnerable.
In present activities, the student shall develop a CAN Intrusion Detection System (IDS) based on Neural Network plausibility data to improve today’s authentication and integrity mechanism system without affecting hardware crypto accelerations resources.
Required skills Neural Networks
Notes In collaboration with PUNCH Softronix
Deadline 16/12/2023
PROPONI LA TUA CANDIDATURA