Research And Implementation Of User Behavior Identification Technology Based On Smartphone

User behavior recognition technology is widely used in identification, production control and human-computer interaction, etc. Traditional user behavior recognition research mostly uses the wearable sensors or external sensors. These sensors are too exoensive and not conducive to marketing. As the rising popularity of smartphones and development of smartphone sensor technology, the thesis proposed the research and implication of user behavior recognition technology based on smartphone.The thesis uses the acceleration sensor of smartphone to collect the signals of standing, walking, running, go upstairs and downstairs. Remove the high frequency noise in the original signal by data preprocessing module, then divide the signals into samples with a length of 512. In the feature extraction and feature selection module, we calculate the mean, standard deviation, correlation coefficient and kurtosis of the acceleration signal in the X axis, Y axis and Z axis to make up of 12 dimension feature vector as the characteristics of the acceleration signal. The thesis uses the support vector machine as a user behavior recognition classifier, and select the 80 of 100 feature vectors to traine it. The rest of the feature vectors are used to test the trained SVM classifier.The experimental results show that the overall recognition rate reached 90.4%, and the standing recognition rate is 100%, walking and running recognition rate are more than 90%, but the recognition rate of go upstairs and downstairs are lower than 85%. In this thesis, the overall recognition rate reached 90.4% by using only one acceleration sensor, it is proved that our scheme is feasible. The main research productions are as follows:1、 The Thesis divided the acceleration signal into fixed length of samples byusing the method of add window.2、 The Thesis proposed a scheme of user behavior recognition based onsmartphone.3、 The Thesis designed and developed a set of user activity recording systembased on the smartphone accelerometer.