Anomaly detection for autonomous driving vehicles

• المؤلف الرئيسي: Chen, Kuilin
• مؤلفون آخرون: Wang Dan Wei
• التنسيق: Thesis-Master by Coursework
• اللغة: English
• منشور في: Nanyang Technological University 2022
• الموضوعات: Engineering::Electrical and electronic engineering::Control and instrumentation::Robotics
• الوصول للمادة أونلاين: https://hdl.handle.net/10356/155805
• المؤسسة: Nanyang Technological University

Autonomous vehicles (AVs) are regarded as the ultimate solution of some social and environmental issues such as traffic accidents, congestion, energy consumption, and emissions. However, it faces more threat of cyber-attack with the increased level of automation because of more complex interior communication in the system and external communication with environment. Although, researchers have done a lot of work on anomaly detection, the validation of these algorithm is expensive and dangerous. The simulation platform, Carla, is developed for autonomous driving development and validation with good performance. In this project, Carla was used to develop anomaly detection model and test working with ROS and Autoware. Five different cyber-attack methods and two anomaly detection approaches were developed. The performance of machine learning based approach, autoencoder, was better than that of model-based approach on the cumulative sum (CUSUM) of residual for some types of attack in a certain driving condition in the experiment.