|LONGARI STEFANO||Cycle: XXXIII |
Section: Computer Science and Engineering
Tutor: PRADELLA MATTEO
Advisor: ZANERO STEFANO Major Research topic
:Assessment of Automotive Attacks and Related Intrusion Detection SystemsAbstract:
Since the early 2000’ the automotive environment has seen great development, the majority of which has been focused on electronics rather than mechanics and material designs. We reached a point where a vehicle cannot be considered only as a mechanical instrument but necessarily also as a network of connected devices. In the last years some of these devices have become so powerful that they can compete in performances with many personal computers, and have been in the meantime connected to the outside world through a plethora of different wireless and wired technologies. Although the increase in electronics and informatics strongly pushes towards higher safety and higher comfort in vehicles, it also brings its own negative effects, and in particular it lowers the intrinsic security of a highly computerized vehicle. In fact, many attacks towards vehicles have been proven possible in the last ten years that raise the question of how to secure such computers on wheels.
The topic of automotive security is obviously being considered by automotive manufacturers, but it is evident how the majority of solutions that are being currently applied are mainly adaptations of more traditional solutions to a new framework rather than specifically tailored solutions designed from scratch for the automotive environment. Although the academia has proposed many different solutions, some more specifically tailored than others, one of the main issues that come as evident while analyzing them is that since automotive manufacturers still apply security-by-obscurity, it is often not completely clear whether such solutions are applicable due to the lack of ground truths. This is particularly evident in the analysis of Intrusion Detection Systems where the ground truths and the attacks have often to be designed and created by the researchers, instead of using real world examples.
Our goal is that of creating a comprehensive analysis of CAN attacks and related intrusion detection systems. We will focus on understanding the strengths of attacks to CAN networks, and the peculiarities of the different possible attacks. This achieved, for all the attacks we focus on understanding the current state of the art in intrusion detection systems and, in case there’s a lack of effective solutions for a given attack, we focus on understanding whether an intrusion detection system that defends from such attack is feasible. If this is not the case we will fall back towards understanding whether there are other security solutions that can mitigate it. At the same time, since as explained the academia still has issues gathering information and in particular well-designed datasets for intrusion detection systems, we aim at gathering the necessary data and structure it so that it can be helpful both to our subsequent analyses both to the academic community as a whole.