Research On Activity Recognition Based On Mobile Phone Acceleration Data

With the development of science and technology,mobile devices are becoming increasingly sophisticated and the latest generation of smart mobile phones now incorporates many diverse and powerful sensors,including the GPS,the vision sensor,the audio sensor,the light sensor,temperature sensors,and the acceleration sensors.The availability of these sensors in mass-marketed communication devices creates exciting new opportunities for data mining oriented applications.This thesis presents a strategy that uses phone-based accelerometers to perform the activity recognition of human body,a task which involves identifying the physical activity a user is behaving.To address this issue,firstly,we collect the acceleration data of some common gestures of daily life by using mobile acceleration sensors,such as walking,jogging,going downstairs,sitting and standing.Then,the data are analyzed to find out the characteristics of human behavior.After that,the raw data is processed,and an effective activity recognition model is established based on classification technology.Finally,the model is used to train and predict the typical human activities,and the results are evaluated and analyzed.This work of human body activity recognition based on the inertial data of mobile phone's acceleration is very beneficial,because the activity recognition model permits us to gain useful knowledge about the habits of millions of users passively——just by letting people carry cell phones in their pockets.This work has a wide range of applications,including the generation of a daily/weekly activity profile to determine if a user(perhaps an obsess child)is performing a healthy amount of exercise.