Priest: Adversarial Attack Detection Techniques for Signal Injection Attacks

Machine learning is widely used for autonomous driving because it can recognize surrounding circumstances feasibly from sensor and determine appropriate actions. Most of these sensors are based on micro-electro-mechanical systems (MEMS), which enable vehicles to judge objects in conjunction with object-detection algorithms. However, recent studies have shown that MEMS vulnerable signal-injection attacks, the input images manipulated force object detection algorithms misclassify results. These attacks be critical wild they deteriorate state-of-the-art techniques, dropping their rates until would no longer detected at all. In this paper, we propose Priest, a novel method against prior attacks. Priest uses similarity pixel values between two consecutive images. Using only ensures low computational cost. According our performance analysis, detects real-time 99% accuracy average, achieving practical security.