Research And Implementation About Device Fingerprints Based On Smartphone's Battery

The widely used device fingerprinting technologies in smartphones are based on some inherent smartphone features,such as IMEI,MAC address and serial number,etc.Besides,some hidden features in devices can also be applied to device identification,such as browser fingerprinting technology(e.g.,fonts,plug-ins,OS version,HTTP headers,and Canvas)and hardware fingerprinting technology(e.g.,CPU and build-in sensors).However,in general,because of the immobilization and stability of the browser fingerprint technology and the dominant feature of the smartphones,the use of them to identify the equipment will bring about different degrees of privacy disclosure.Meanwhile,the hardware fingerprint technology needs to be calibrated before being.used,and the accuracy of calibration greatly affects the final recognition accuracy,which leads to a reduction in practicability.Android system developers are also constantly upgrading the system,which leads to thedecrease of the accuracy of some of the original schemes,so it is a surprisingly challenging task to find a new,secure and effective smartphone fingerprint.A principled approach to its design is therefore,either to identify smartphones effectively,and to protect user' smartphones through simple experimental operation.In this paper,a new battery based hardware fingerprint identification scheme isproposed,which uses the characteristics of smartphones' power consumption to identify smartphones.The discrepancies on manufacturing of smartphones make the power consumption is different when performs the same task.In order to get more power consumption characteristics,we stimulate the algorithm with different power consumption of tasks.Moreover,power consumption of smartphones can be easily obtained without strict operating steps.We use some tools to collect the energy consumption information,and use the supervised learning method to classify the smartphones by using the extracted power consumption characteristics.Finally,we use 15 smartphones in both laboratory and public conditions toevaluate the performance of the battery fingerprinting from various angles.The experimental results indicate that battery fingerprint can be efficiently used to identify the smartphones with low overhead.At the same time,it will not bring privacy problems,since the power consumption information is changing in real time.