Machine Learning Approaches for WiFi Round Trip Time Indoor Positioning Systems - Doctoral Dissertation

Aug 5, 2026

Upload by Levi

Indoor positioning systems using WiFi Round-Trip Time (RTT) measurements under IEEE 802.11mc can reach sub-metre accuracy via trilateration in ideal line-of-sight conditions, yet performance in complex non-line-of-sight (NLOS) settings remains uncertain. This dissertation develops new machine learning algorithms and validates them with extensive real-world experiments. It analyzes WiFi RTT properties across heterogeneous smartphones and releases three public datasets including RTT and RSS signals. The work presents an NLOS identification feature selection method reaching up to 98% accuracy and a weighted model-switching approach that improves positioning by up to 1.8 metres over standard fingerprinting.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages216

Reading Time9 h 4 min

AccessFree to download

Abstract

WiFi RTT overview and research motivation

Proposed datasets and baseline results

NLOS identification via feature selection

Dynamic switching of positioning models

Machine Learning Approaches for WiFi Round Trip Time Indoor Positioning Systems Xu Feng Computer Science Department Royal Holloway University of London This dissertation is submitted for the degree of Doctor of Philosophy February 2025 To those I love and who love me most, and especially to my dear younger self, for whom I have become the strength, love, and support I once needed. Declaration I hereby declare that except where specific reference is made to the work of others, the contents of this dissertation are original and have not been submitted in whole or in part for consideration for any other degree or qualification in this, or any other university. This dissertation is my own work and contains nothing which is the outcome of work done in collaboration with others, except as specified in the text and Acknowledgements. Xu Feng February 2025 Acknowledgements Foremost, I want to express my deepest gratitude to my supervisors, Dr. Khuong An Nguyen and Prof. Zhiyuan Luo, for their invaluable guidance, expertise, and encouragement over the past four years. Their thoughtful feedback and insightful suggestions have been instrumental in shaping this work. Beyond their academic mentorship, I am profoundly grateful for their patience, kindness, generosity and consistent support, which created a nurturing environment that allowed me to overcome challenges and grow. Their dedication to nurturing my growth and inspiring my pursuit of excellence has been a constant source of motivation. I am truly fortunate to have had their time, knowledge, and expertise, as their guidance and influence have been invaluable throughout my journey. I also wish to extend my heartfelt thanks to the Computer Science Department for creating a warm and supportive environment. The resources, facilities, and opportunities provided by the department have played a significant role in the success of this research. It has been a privilege to be part of such a vibrant and inspiring academic community. I am deeply grateful to Prof. Guang Li of Zhejiang University, Liudi, and Huhu (my cat), whose unwavering belief in me gave me the confidence to start this journey. You were my earliest motivators, the bedrock of my aspirations, and the strongest pillars of support during this transformative chapter. Your encouragement has been the foundation of this achievement, and I cannot thank you enough. To my dear friends Weiwei, Xingyu, and Gina, thank you for your kindness, encouragement, and the moments of joy and laughter that brought balance to my life. Your friendship has been a true blessing, providing support through challenges and celebrating the joy of victories along the journey. Finally, to my family, I owe my deepest gratitude for their unconditional love, encouragement, and faith in me. I am especially thankful to my family members Dongqin, Qing, Jianxue, and Lanhua, whose steadfast support and love sustained me through countless late nights. Your presence in my life has been my greatest strength. This thesis is not only a reflection of my effort but also a testament to the collective support and encouragement I have received from all of you. Thank you for being a part of this journey. Abstract Indoor positioning systems based on WiFi Round-Trip Time (RTT) measurements, as per the IEEE 802.11mc standard, have demonstrated sub-metre level accuracy using trilateration under ideal indoor conditions. However, the efficacy of WiFi RTT positioning in complex, non-line-of-sight (NLOS) environments remains an open research question. Therefore, this thesis addresses the challenge by proposing novel machine learning algorithms and validating their performance through extensive empirical experiments in real-world testbeds. Recent literature has shown improvements in WiFi fingerprinting systems utilising deep learning methods, achieving sub-metre accuracy. However, it was observed that simpler neural networks can sometimes outperform comp

FAQs

Why is WiFi RTT indoor positioning challenging in real environments?

While WiFi RTT supports sub-metre accuracy under line-of-sight conditions, typical workplaces introduce reflection, attenuation, and diffraction that lead to non-line-of-sight (NLOS) effects, reducing reliability.

What datasets are introduced in the dissertation?

The dissertation provides three publicly available datasets collected from large-scale real-world scenarios. They include both RTT and received signal strength (RSS) measurements gathered on heterogeneous smartphones.

How does the dissertation detect NLOS conditions for WiFi access points?

It proposes a feature selection algorithm using RSS and RTT as inputs, employing multi-scale selection and machine-learning-based weighting. The method achieves up to 98% accuracy in NLOS detection of access points.

Aug 5, 2026

Upload by Levi