Live Aircraft Destination Prediction using Machine Learning & Adversarial Attacks

Abstract

As a continuation of uncovering threats posed in aviation that are due to ADS-B not being secure, we looked into how a non-commercial aircraft user's privacy can be compromised in real-time with publicly available data. Since there are many ways to predict the destination of an airplane whilst in the air we investigated this matter using a long short-term memory and a gradient booster model. With the gradient booster model, depending on the planes operator, we successfully achieved an accuracy of 0.70 to 0.89, 30 minutes before the aircraft lands. In addition, with adversarial machine learning attacks, we deluded the gradient booster model, while keeping the attacker's capabilities as omnipotent but still achievable as possible. We managed to misdirect the model with a naive based and targeted based poisoning attack. The most effective attack on the model seemed to be the targeted one, it had the greatest dissimulation with the least data manipulation.