Using Smartphones for Enhanced Integrated System Strategies for Indoor Pedestrian Positioning

Abstract

In recent years smartphones have become nearly ubiquitous Due to the increased accuracy and decreased cost of Micro-electro mechanical system (MEMS) sensors the smartphone has come to be embedded with various sensors which improve its usefulness for various mobile applications These have led to excitement over the future possibilities of Location-Based Services (LBS) and the Internet of Things (IoT) The core component of LBS and IoT is the position of the user or device causing renewed attention to be given to twenty-year old positioning technologies related to mobile and indoor positioning The accuracy and stability of Global Navigation Satellite Systems (GNSS) for outdoor positioning is already reliable and mature However indoors the blocked of satellite signals make it difficult to obtain accurate and continuous positioning Many kinds of mature indoor positioning technologies have been proposed in the past but due to the different background of the past most of these require additional infrastructure and wearable sensors rather than exploiting the latent usefulness of smartphones with their built-in low-cost sensors It is true that this technology has limitations such as the lack of real-time ability because of the need of prior knowledge and posterior modeling; the costs of additional infrastructure and devices; the tiresome need for successively taking images or applying expert knowledge; the time-consuming creation of an environmental database (which can suddenly become inaccurate due to environmental changes); computation-heavy process requiring high-grade hardware; and the rapid accumulation of error caused by inaccurate sensors In the other words due to all of these shortcomings at the present time no indoor positioning technology has an overwhelming advantages that meets the requirements of human life and industrial development Therefore although researchers have been working on indoor positioning technologies for many years they are still trying to overcome these challenges and design an indoor positioning system which can be widely used in human life Put another way a single positioning system can be seen as a double-edged sword When the system utilizes its positioning principles and corresponding measurements the inherent shortcomings are usually unavoidable Therefore integration and customizability are possible solutions at this stage: by using different indoor positioning technologies in coordination their advantages and shortcomings can be made complementary In other words choosing the appropriate technology depends on the environment for which the integrated positioning system can be customized to achieve accurate low-cost indoor positioning Geomatics can provide another approach for developing this technology by employing coordinate systems an accurate map and floor plan as well as positioning technology originally set up for surveying and mapping In considering these approaches to the problem of indoor navigation the author has developed some enhanced strategies using various positioning technologies not only employing mainstream knowledge but also cutting-edge Geomatics After applying these systems the integrations and algorithms are verified and evaluated at various test sites To begin this study proposes the use of integrated architecture for indoor positioning based on the sensors embedded in smartphones as well as other common positioning technologies with the core approach based on inertial positioning technology To verify this different integrations and enhanced algorithms have been developed for this study’s integrated architecture: 1) walking velocity constraint and real-time smoothing for pedestrian inertial navigation; 2) map-aided pedestrian dead reckoning (PDR) using a fuzzy decision tree; 3) space resection aided PDR and integration based on PDR and ANN aided marker-based image localization; 3) a differential Bluetooth positioning system As preliminary verification these strategies were evaluated for this study and found to be performing well boding well for the success of the proposed ultimate integrated architecture However since there is a lot of related technology in this complex architectural system numerous tasks await later research projects The proposed integrated architecture can only be considered a truly excellent indoor positioning application after all of the remaining parts are finished and integrated In addition the specifications and performance of the sensors which were employed are important for developing corresponding technologies as is the map floor plan and georeferenced images used in this study Therefore the calibrations of the sensors and the mobile mapping procedures are attached and the experience of developing an indoor positioning system in a real site is presented The purpose of this thesis is not only to make a technological contribution but also to serve as a guide for those who which to work in the field of sensing and positioning

Date of Award	2018 Jan 18
Original language	English
Supervisor	Kai-Wei Chiang (Supervisor)