Multi-sensor systems and models for high accuracy indoor positioning
File(s)
Author(s)
Constantinou, Petrina
Type
Thesis
Abstract
Technological advances over the years, gave rise to location based services (LBS) and their demand in terms of higher performance, measured by the required navigation performance parameters including accuracy and availability. Until recently this has been prominent in outdoor environments, where global navigation satellite systems (GNSS) facilitated outdoor LBS, triggering the rise in their popularity and making GNSS a de facto positioning technology in outdoor environments. Advances in position navigation and timing systems, and communication technologies, are gradually increasing the potential of high performance indoor positioning systems (IPS), increasing the demand for more indoor LBS. The complexity of indoor environments, presents serious challenges to IPS, and currently there is no consensus on a generic positioning technology for indoor environments.
Currently the approach taken involves identifying applications, specification of requirements and architectural design. However, a critical consideration of IPS design should be a detailed understanding of the ’operations’. This is to enable a complete identification of functions that require positioning.
Therefore, this thesis firstly takes the approach that identifies indoor applications and
their requirements. It then specifies in detail the operational functional system architecture of a cognitive agent, and identifies these functions that require positioning within them, the location based functions (LBF). This enables confirmation and augmentation where necessary to meet the applications and their requirements. The thesis then explores the technologies required to implement LBF (i.e. physical architecture). Candidate technologies are reviewed and based on accuracy performance, UWB and INS are analysed individually and in integration. A UWB/INS integration using UWB-derived speed estimates, measurement corrections and filtering is developed. The results show the proposed integration improves accuracy (95%) performance up to 65.5% compared to UWB-standalone system and 36.6% compared to UWB/INS integration without UWB-derived speed estimates.
Currently the approach taken involves identifying applications, specification of requirements and architectural design. However, a critical consideration of IPS design should be a detailed understanding of the ’operations’. This is to enable a complete identification of functions that require positioning.
Therefore, this thesis firstly takes the approach that identifies indoor applications and
their requirements. It then specifies in detail the operational functional system architecture of a cognitive agent, and identifies these functions that require positioning within them, the location based functions (LBF). This enables confirmation and augmentation where necessary to meet the applications and their requirements. The thesis then explores the technologies required to implement LBF (i.e. physical architecture). Candidate technologies are reviewed and based on accuracy performance, UWB and INS are analysed individually and in integration. A UWB/INS integration using UWB-derived speed estimates, measurement corrections and filtering is developed. The results show the proposed integration improves accuracy (95%) performance up to 65.5% compared to UWB-standalone system and 36.6% compared to UWB/INS integration without UWB-derived speed estimates.
Version
Open Access
Date Issued
2024-02-29
Date Awarded
01/12/2024
License URL
Advisor
Ochieng, Washington Yotto
Feng, Felix
Sponsor
Engineering and Physical Sciences Research Council
Publisher Department
Civil and Environmental Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)