Implementation of Smartphone-based Pedestrian Dead Reckoning for Indoor Positioning Systems

• Main Authors: SHIH-MIN HUANG, 黃仕旻
• Other Authors: 李俊賢
• Format: Others
• Online Access: http://ndltd.ncl.edu.tw/handle/sm92fm

碩士 === 國立臺北科技大學 === 電機工程研究所 === 105 === The large-scale and complex-structure buildings are on increase in the recent year, such as airports, department stores, and train stations. In order to make people more convenient and efficient to reach the destination in a large indoor place, indoor localization techniques get more and more attention. There are many indoor localization methods currently, such as triangulation, fingerprinting, and pedestrian dead reckoning (PDR) and so on. The method about PDR is mounting the inertial measurement unit (IMU) on some parts of the body or using sensors inside the smartphone to analyze the signal when people walk, and then further calculate the distance and direction of movement to compute user’s position. However, when using the PDR method, users should carry the smartphone at one part of the body, which will cause some restrictions. In order to achieve the effects of cost and convenient using, this paper uses the smartphone together with PDR method for indoor localization. This paper proposes the method of heading determination to reduce the cumulative error of the angle. Four modes of carrying phone position, including holding, pocket, swing, and calling are taken into consideration in order to solve the problem of the phone carried in a fixed position. Depending on the different modes, the proposed system will use the different methods of step detection and heading determination while simultaneously estimating each step length. The result got from such procedure is then combined with an indoor map, and users can get positioning information. Moreover, the proposed system can classify the modes of phone position accurately, and also achieve localization performance with an average position accuracy of 96.63% in the four modes.