Activity recognition from smartphone sensing data

Recognizing human activities with sensors next to the body has become more important over the years, aiming to create or improve systems in elder care support, health/fitness monitoring, and assisting those with cognitive disorders. This task of recognizing activities taking place at a certain moment when considering only one individual user is called Activity Recognition. With smartphones, it was possible to overcome the barrier of the person having to use multiple body worn sensors. Having the smartphones a triaxial accelerometer built in it is possible to create applications that are capable of recognizing the activities of the user with great accuracy. This thesis aims to meet this new solution, creating an Android application, for the problem of recognizing the activities performed by the user and treating it as a classification problem. This work should be useful for anyone who studies the human body and should help to improve the user experience with is phone.