Wi-Fi-based localization for indoor navigation

• Main Author: Ng, Ze Wei
• Other Authors: Xie Lihua
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Electrical and electronic engineering::Wireless communication systems; Engineering::Electrical and electronic engineering::Control and instrumentation::Robotics
• Online Access: https://hdl.handle.net/10356/172221
• Institution: Nanyang Technological University

Being able to localize oneself with respect to an environment is an important aspect of numerous industries such as robotics or autonomous vehicles. However, the traditional method of using Global Positioning System (GPS) to localize may pose a challenge in several use cases where these signals are unable to reach the user. This study proposes a novel design of using ESP32 microcontrollers to facilitate localization in environments with poor GPS access – but with the presence of Wi-Fi access points of known position. This report presents a proof of concept leveraging vSLAM for autonomous data collection, proposes several machine learning and deep learning models to achieve Wi-Fi based localization, and illustrates an end-to-end architecture for inference and frontend display of pose estimates against a map.