Detecting Bluetooth Attacks Against Smartphones by Device Status Recognition

Abstract

Bluetooth is a universal wireless standard which is used on most smartphones. With the widespread use of Bluetooth on smartphones, Bluetooth security has received a lot of attention, and it is increasingly important to identify and block Bluetooth attacks to ensure that smartphones are free of the threat of Bluetooth attacks. Traditional attack detection techniques are generally based on traffic, and unfortunately smartphone Bluetooth traffic is extremely difficult to capture. Based on this problem, we propose a Bluetooth attack detection method based on device status recognition, which uses the response time of the smartphone to the ping in the Bluetooth L2CAP to remotely monitor the Bluetooth status of the smartphone. Due to the variety of Bluetooth attacks, we can’t easily identify multiple Bluetooth attacks based on response time. For this purpose, we explored the effectiveness of a detection approach based on deep learning. Our experimental results show our algorithm can detect Bluetooth attack with a high precision and high recall.