Overlapped-shaking: A local authentication method for mobile applications

Abstract

Mobile applications become increasingly popular currently. Both authentication and privacy are very important for mobile applications. A promising approach for local authentication is to match the acceleration sensor samples of the two candidate smart phones. However, the acceleration sensor on the smart phones sample at different rates and the starting phases are not aligned. This paper presents a new local authentication method for mobile applications called Overlapped-shaking. Overlapped-shaking is a background service for smart phones which can measure the similarities between two shaking patterns. We design a new acceleration samples matching method called SAM, which can handle the data in different sample rates and offset at the beginning. We implement a demonstration of Overlapped-shaking service on Android smart phones. Test results show that our approach works very well for acceleration-based smart phone authentications.