Motion and audio analysis in mobile devices for remote monitoring of physical activities and user authentication

摘要

In this article we propose the use of accelerometer embedded by default in smartphone as a cost-effective, reliable and efficient way to provide remote physical activity monitoring for the elderly and people requiring healthcare service. Mobile phones are regularly carried by users during their day-to-day work routine, physical movement information can be captured by the mobile phone accelerometer, processed and sent to a remote server for monitoring. The acceleration pattern can deliver information related to the pattern of physical activities the user is engaged in. We further show how this technique can be extended to provide implicit real-time security by analysing unexpected movements captured by the phone accelerometer, and automatically locking the phone in such situation to prevent unauthorised access. This technique is also shown to provide implicit continuous user authentication, by capturing regular user movements such as walking, and requesting for re-authentication whenever it detects a non-regular movement.