Design Of Action Recognition Based Mobile Map Mavigation System

With the popularity mobile internet, electronic maps and other location-based applications emerged as the most promising application forms. In location-based applications, it is an urgent problem to obtain accurate position of users. There are three main methods to obtain users’ position, including GPS, GSM positioning. Studies show that GSP is shortcomings of too much power consumption, and GSM positioning has low positioning accuracy. Most smart phones embedded acceleration sensors, which makes it possible to apply wearable computing to mobile internet. Different body movements often have different speed, and different speed can be used as a reference to estimate the displacement of body, so action recognition-based location is possible.An action recognition based mobile map which is called TuShou map in Android is designed and developed in this paper. Firstly, current situation of mobile maps and TuShou map is introduced. After that, inertial positioning and GPS, GSM hybrid positioning is used in this system. First, the requirements and core processes of activity recognition based mobile map navigation system are introduced. Then, activity recognition subsystem, hybrid positioning subsystem and nabigation subsystem are introduced. Activity recognition subsystem is introduced exhaustively. Traveling speed and displacement of body is estimated through current human body. After that hybrid positioning subsystem of displacement estimation and GPS, GSM is designed.