الفهرس 
State of Art
Literature SurveyOnly 14 pages are availabe for public view
 |  | from | 136 |  |  |
| | | from | 136 | | |
Abstract
Security nowadays is essential for internet-connected systems that we use daily and may contain our sensitive data. In the sense of computing, security is composed of cybersecurity and physical security. Cybersecurity protects the inter-connected networks, including software, hardware, and data from cyberattacks. While, physical security is the protection of hardware and software from illegal physical acts that may cause serious damage in any enterprise or organization.
Cyber-physical systems are physical and computational process collaboration. Modern automobiles can be thought of as a cyber-physical networked system. For example, autonomous driving requires a strong vehicle interconnection and an opening up to external information sources and services, therefore increasing the potential surface of the attack. The electronic control unit (ECU), which is a group of computing devices, has recently achieved considerable advancements in the automobile environment. To increase energy efficiency and lower noise and vibration, an ECU is utilised to track and control the subsystems of a vehicle. Practical experiments have already shown that an intruder could obtain remote access to an ECU in the car.
Many cyber-physical systems, such as vehicles are vulnerable to many attacks as a result of their automation and the trials to digitalize them every day. The mitigation and detection of these attacks have been increased recently due to its importance. The idea of this research is to design and implement an intrusion detection system (IDS) that can identify the cyber-physical attacks for vehicles using one or more security techniques.
We propose three frameworks where each one detects a certain attack model. Two datasets, which are Seat Leon 2018 and KIA SOUL, are used to evaluate the three frameworks. The experimental results of the first framework show that it can detect basic point attacks of the first attack model with an accuracy of 94.00% for the Seat Leon 2018 dataset and 99.13% for the KIA SOUL dataset. While the experimental results of the second framework show that it can detect more complex point abnormalities of the second attack model with an accuracy of 97.00% for the Seat Leon 2018 dataset and 97.49% for the KIA SOUL dataset. The results of the third framework against different point and period attacks show the superiority of the framework and its robustness with high accuracy of 99.05% for Seat Leon 2018 and an accuracy of 99.22% for KIA SOUL.