Detection of attacks and intrusions on automotive engine IoT sensors

Abstract

Predictive maintenance is used to predict system failures using deep learning algorithms and IoT sensors. However, IoT sensors and deep learning algorithms are susceptible to attacks, which at the same time poses a serious threat as far as car engine IoT sensors are concerned. This paper tends to research the consequence of false data injection on IoT automotive engine sensors, which can result in disastrous results. Also, the following deep learning algorithms are used in this paper to detect attacks and intrusions on automotive engine IoT sensors: RNN (Recurrent Neural Networks), LSTM (Long Short Term Memory Networks), GAN (Generative Adversarial Networks) and a new developed algorithm SPNN (Sequential Probability Neural Networks). The new SPNN algorithm was the fastest in detecting and preventing attacks/intrusions on automotive engine IoT sensors when it came to continuous attack, but the GAN algorithm was the fastest in detecting and preventing attacks/intrusions on automotive engine IoT sensors when it came to temporary attack.