Authentication of Smartphone Users Based on Activity Recognition and Mobile Sensing

Abstract

Smartphones are context-aware devices that provide a compelling platform for ubiquitous computing and assist users in accomplishing many of their routine tasks anytime and anywhere, such as sending and receiving emails. The nature of tasks conducted with these devices has evolved with the exponential increase in the sensing and computing capabilities of a smartphone. Due to the ease of use and convenience, many users tend to store their private data, such as personal identifiers and bank account details, on their smartphone. However, this sensitive data can be vulnerable if the device gets stolen or lost. A traditional approach for protecting this type of data on mobile devices is to authenticate users with mechanisms such as PINs, passwords, and fingerprint recognition. However, these techniques are vulnerable to user compliance and a plethora of attacks, such as smudge attacks. The work in this paper addresses these challenges by proposing a novel authentication framework, which is based on recognizing the behavioral traits of smartphone users using the embedded sensors of smartphone, such as Accelerometer, Gyroscope and Magnetometer. The proposed framework also provides a platform for carrying out multi-class smart user authentication, which provides different levels of access to a wide range of smartphone users. This work has been validated with a series of experiments, which demonstrate the effectiveness of the proposed framework.

Context-Aware Human Activity Recognition (CAHAR) in-the-Wild Using Smartphone Accelerometer

“…As the proposed CAHAR scheme aims to effectively discriminate between the varying physical activity patterns in different contexts, hence it is necessary to choose a robust set of features for activity classification. For this purpose, we extracted eighteen (18) time-domain features from the preprocessed accelerometer data, which have revealed efficient performance in HAR related studies [10], [11], [35]. These features provide essential signal traits that are significant in recognizing human activities based on sensor data.…”

Securing PIN‐based authentication in smartwatches with just two gestures

“…However, tests were performed in a lab with a limited dataset, which may favor the classification process. All works presented above are related to the specific behavior of a human while performing some tasks, such as hand movement, gait, and rhythmic tapping, which may present some limitations . For example, multiple users may have the same hand‐waving patterns.…”

A Continuous Identity Authentication Scheme Based on Physiological and Behavioral Characteristics

“…Research on authentication by motion sensors is relatively new. Recently, with the increasing capability of smartphones, Ehatishamulhaq et al [ 14 ] used the embedded motion sensors of smartphone for users’ authentication. They applied several classifiers to recognize different activities, then authenticated the identity of a user based on the prior knowledge of their motion states.…”