Design And Implementation Of Mobile Phone Pedometer Based On Android Sensor Module

In recent years,sensors have been widely applied on smartphones as the basic devices of information acquisition,which can achieve auto adjust-brightness,auto rotate-screen,intelligent mistake-precation touch,motion simulation,so as to make smartphone's functions more and more abundant and make human-computer interactive experiences more and more friendly.As an important part of the smartphone,the sensor module has become a hot research topic in recent years.One of the reasons is that various basic functions of sensors are realized,the other reason is that using different sensors' data rationally can also achieve human motion monitoring functions such as pedometer,fall detection,behavior recognition.Especially mobile phone pedometer,as one application instance of sensor module,can conveniently monitor user's daily movement.In view of the significant contribution that sensors have made for human-computer interaction experiences and the human pressing needs for movement monitoring,the Android sensor module is designed and implemented at frist,and then pedometer funtion is implemented on the basis of it in this thesis.The main contents are divided into the following three parts: first,an multi-sensor fusion software architecture is designed in detail base on multi-sensors data fusion scheme which is proposed by STMicroelectronics;Second,according to the design of the software framework,the Android driver layer and hardware abstraction layer of accelerometer and proximity sensor are designed and implemented,including load and unload of driver,acquisition and report of sensor data,sensor calibration module,etc.A fast calibration algorithm is presented to improve the reliability of proximity sensor data;Third,the human motion model is established,and accelerometer data are collected,pretreated and filtered when user is walking,running,walking up and down the stairs,and finally a pedometer algorithm that fuse data of accelerometer and proximity sensor is designed.The algorithm proposes a method to filter interferential peak of accelerometer data through amplitude threshold and time threshold,to add proximity sensor data to judge smartphone position,then completes step detection through detecting effective peak,improves the lower pedometer precision when users hold phones.At the end of this thesis,sensor module and pedometer module are tested and analyzed.Sensor module tests mainly include functional test and data comparative test before and after the calibration.The test results show that the accelerometer and proximity sensor acquire and report data successfully and improve data reliability after calibration;Pedometer module tests mainly include pedometer precision test and comparative test of different movements.The test results show that this pedometer has higher precision when users hold phones and has stronger anti-interference ability in some non-step-detection scenes such as taking a bus or car.The average precision of this pedometer algorithm is more than 95% in walking,more than 96% in walking up and down the stairs,more than 98% in running,which all meet the design requirements.