DMRC

Software Engineering for AI project's repo

Evasion attacks on self-driving car

Introduction

The project concerns possible evasion attacks on traffic sign recognition on self-driving cars. We used CARLA simulator to acquire the dataset and then trained different ML models to classify the images. Finally, additional datasets were created with evasion attacks applied to be tested on the previously trained models. The last step was to evaluate the resilience level of the different models with the various attacks .

Contributors