An Autoencoder Based Approach to Defend Against Adversarial Attacks for Autonomous Vehicles

Boosted by the evolution of machine learning technology, large amount of data and advanced computing system, neural networks have achieved state-of-the-art performance that even exceeds human capability in many applications [1] [2] . However, adversarial attacks targeting neural networks have demonstrated detrimental impact in autonomous driving [3] . The adversarial attacks are capable of arbitrarily manipulating the neural network classification results with different input data which is non-perceivable to human.