Use this URL to cite or link to this record in EThOS:
Title: RSSI based self-adaptive algorithms targeting indoor localisation under complex non-line of sight environments
Author: Hou, Xiaoyue
ISNI:       0000 0004 7654 8136
Awarding Body: University of Edinburgh
Current Institution: University of Edinburgh
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Location Based Services (LBS) are a relatively recent multidisciplinary field which brings together many aspects of the fields of hardware design, digital signal processing (DSP), digital image processing (DIP), algorithm design in mathematics, and systematic implementation. LBS provide indirect location information from a variety of sensors and present these in an understandable and intuitive way to users by employing theories of data science and deep learning. Indoor positioning, which is one of the sub-applications of LBS, has become increasingly important with the development of sensor techniques and smart algorithms. The aim of this thesis is to explore the utilisation of indoor positioning algorithms under complex Non-Line of sight (LOS) environments in order to meet the requirements of both commercial and civil indoor localisation services. This thesis presents specific designs and implementations of solutions for indoor positioning systems from signal processing to positioning algorithms. Recently, with the advent of the protocol for the Bluetooth 4.0 technique, which is also called Bluetooth Low Energy (BLE), researchers have increasingly begun to focus on developing received signal strength (RSS) based indoor localisation systems, as BLE based indoor positioning systems boast the advantages of lower cost and easier deployment condition. At the meantime, information providers of indoor positioning systems are not limited by RSS based sensors. Accelerometer and magnetic field sensors may also being applied for providing positioning information by referring to the users' motion and orientation. With regards to this, both indoor localisation accuracy and positioning system stability can be increased by using hybrid positioning information sources in which these sensors are utilised in tandem. Whereas both RSS based sensors, such as BLE sensors, and other positioning information providers are limited by the fact that positioning information cannot be observed or acquired directly, which can be summarised into the Hidden Markov Mode (HMM). This work conducts a basic survey of indoor positioning systems, which include localisation platforms, using different hardware and different positioning algorithms based on these positioning platforms. By comparing the advantages of different hardware platforms and their corresponding algorithms, a Received Signal Strength Indicator (RSSI) based positioning technique using BLE is selected as the main carrier of the proposed positioning systems in this research. The transmission characteristics of BLE signals are then introduced, and the basic theory of indoor transmission modes is detailed. Two filters, the smooth filter and the wavelet filter are utilised to de-noise the RSSI sequence in order to increase localisation accuracy. The theory behind these two filter types is introduced, and a set of experiments are conducted to compare the performance of these filters. The utilisation of two positioning systems is then introduced. A novel, off-set centroid core localisation algorithm is proposed firstly and the second one is a modified Monte Carlo localisation (MCL) algorithm based system. The first positioning algorithm utilises BLE as a positioning information provider and is implemented with a weighted framework for increasing localisation accuracy and system stability. The MCL algorithm is tailor-made in order to locate users' position in an indoor environment using BLE and data received by sensors locating user position in an indoor environment. The key features in these systems are summarised in the following: the capacity of BLE to compute user position and achieve good adaptability in different environmental conditions, and the compatibility of implementing different information sources into these systems is very high. The contributions of this thesis are as follows: Two different filters were tailor-made for de-nosing the RSSI sequence. By applying these two filters, the localisation error caused by small scale fading is reduced significantly. In addition, the implementation for the two proposed are described. By using the proposed centroid core positioning algorithm in combination with a weighted framework, localisation inaccuracy is no greater than 5 metres under most complex indoor environmental conditions. Furthermore, MCL is modified and tailored for use with BLE and other sensor readings in order to compute user positioning in complex indoor environments. By using sensor readings from BLE beacons and other sensors, the stability and accuracy of the MCL based indoor position system is increased further.
Supervisor: Arslan, Tughrul ; Flynn, Brian Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: GPS ; indoor positioning ; Bluetooth 4.0 beacons ; Monte Carlo Localisation