Wuhan University and Amap Win the Smartphone‑Based Indoor Positioning Track at IPIN2021

In an era where smartphones are ubiquitous, navigation relies on a complex set of technologies, especially positioning, which underpins the user experience of map services. The International Conference on Indoor Positioning and Indoor Navigation (IPIN) 2021, held in Spain, featured a competition where the joint team of Wuhan University and Amap (Gaode Map) secured the champion title in the "Smartphone‑Based Indoor Positioning" track.

IPIN, founded in 2010, is the world’s largest academic conference on indoor positioning and ranks among the top‑tier venues in the field. The 2021 competition comprised three tracks; the indoor positioning track, established in 2015, attracts the most participants and longest history.

The challenge required teams to use only the built‑in sensors of a smartphone—GNSS, inertial sensors, magnetometer, Wi‑Fi, Bluetooth, barometer, and ambient light—to achieve real‑time pedestrian positioning without any additional infrastructure. Participants simulated everyday phone usage (e.g., climbing stairs, making calls, screen input) across multiple floors, and the evaluation used data from a randomly selected device.

Key difficulties included the absence of GNSS signals indoors, prohibition of dedicated positioning hardware (e.g., UWB, Bluetooth AOA), and the need to infer floor level and handle diverse user behaviors, which complicated inertial sensor data.

The winning team designed a fusion framework that fully exploits smartphone sensor capabilities. They combined a machine‑learning‑driven pedestrian dead‑reckoning (PDR) algorithm with magnetic, Bluetooth, and Wi‑Fi signal matching, and applied a filter to merge outputs from all modules, achieving high‑frequency, accurate real‑time positioning.

This approach merged traditional network positioning, inertial navigation, and advanced geomagnetic techniques, leading to a decisive victory in the competition. The technology promises to enhance indoor navigation scenarios such as parking lot vehicle locating, ride‑hailing pick‑up points, and in‑store navigation.

Amap’s long‑standing expertise in satellite, network, inertial, visual, and geomagnetic positioning, combined with big‑data analytics and machine‑learning models, underpins the solution and will continue to drive future innovations in precise, convenient travel experiences.