Detection of smart cars cyber attacks
| Title | Detection of smart cars cyber attacks |
|---|---|
| Summary | For treating the probem of cyber attacks against smart vehicles, new change-point detection and anomaly detection methods by means of statistics and machine learning are developed and evaluated. |
| Keywords | Smart vehicles, Cloud security, Change-point detection, Machine learning, Performance measures |
| TimeFrame | Spring 2020 |
| References | Weber et al: Embedded Hybrid Anomaly Detection for Automotive
CAN Communication, Weber et al: Online Detection of Anomalies in Vehicle Signals using Replicator Neural Networks |
| Prerequisites | Mathematical statistics, machine learning, programming |
| Author | |
| Supervisor | Eric Järpe (HH), Cristofer Englund (HH and RISE Viktoria), Ana Magazinius (RISE Viktoria) |
| Level | Master |
| Status | Open |
As the communication and control of cars, buses and trucks is getting smarter, so are the threats from malicious attempts to compromize security of these vehicles. The problem of detecting and possibly classifying such attempts is approached from the classical statistical change-point detection angle but also from the newer machine learning side. Also hybrid methods could be considered depending on how the project proceeds. The goal is to develop new cloud methods and to evaluate the methods in comparison to state-of-the-art methods used today. Possibly, new performance measures could be constructed in order to better understand how well the suggested methods behave for the environment-specific problems. The methods could possibly be demonstrated in examples with real data. The thesis will be a part of the CyRev research project in collaboration with partners from the industry.
