As the automotive industry shifts toward electric and autonomous vehicles, security is more important than ever. Research into the field by two UWindsor engineering students was honoured as best paper at the International Conference on Wireless and Satellite Systems, held in Singapore last month.

Pegah Mansourian and Mina Zamanirafe, doctoral students of computer and electrical engineering, developed machine learning software that can detect an attack on systems before they have an effect.

“We want to ensure they are safe as possible for the consumer,” says Mansourian. “If an attack was to occur while a person is driving it could be very dangerous, as the driver wouldn’t have control.”

Their paper, “Anomaly Detection for Connected Autonomous Vehicles using LSTM and Gaussian Naïve Bayes,” co-authored with engineering professor Ning Zhang, computer science professor Arunita Jaekel, and Marc Kneppers, chief security architect for Telus, telecommunications company, focuses on the development of software for the in-vehicle experience.

Dr. Zhang notes that it proposes innovative intrusion detection solutions based on deep learning.

“Our solutions can enable fast and accurate identification of different types of attacks to mitigate their impact, ensuring the safety of connected and autonomous vehicles and their users,” he says.