Hybrid Wi-Fi/PDR Indoor Localization with Fingerprint Matching
Chunyi Zhang, Zongwei Li, Xiaoqi Li
TL;DR
This paper addresses the challenge of accurate, low-cost indoor localization by fusing Wi-Fi RSS fingerprinting with Pedestrian Dead Reckoning (PDR). It presents a hybrid approach that uses a dynamic weighted WKNN for fingerprint matching and leverages PDR for real-time trajectory updates, implemented on Android with a Spring Boot backend and MySQL database. The key contributions include a distance-based dynamic weighting scheme for WKNN, a complete system architecture, and validation through Matlab simulations and field tests showing stable latency and accuracy. The work demonstrates a practical, scalable framework for deploying indoor positioning in real-world environments without additional hardware beyond existing Wi-Fi infrastructure and commodity smartphones.
Abstract
Indoor position technology has become one of the research highlights in the Internet of Things (IoT), but there is still a lack of universal, low-cost, and high-precision solutions. This paper conducts research on indoor position technology based on location fingerprints and proposes a practical hybrid indoor positioning system. In this experiment, the location fingerprint database is established by using RSS signal in the offline stage, the location algorithm is improved and innovated in the online stage. The weighted k-nearest neighbor algorithm is used for location fingerprint matching and pedestrian dead reckoning technology is used for trajectory tracking. This paper designs and implements an indoor position system that performs the functions of data collection, positioning, and position tracking. Through the test, it is found that it can meet the requirements of indoor positioning.