Previous Abstract Return to Session B6 Next Abstract

Session B6: Emerging PNT Consumer Applications

Smartphone GNSS Booster: Centimeter-Level Pedestrian Positioning Using a Portable Signal Re-Radiator
Taro Suzuki, Chiba Institute of Technology
Location: Prince David Room
Date/Time: Thursday, Apr. 16, 1:27 p.m.

The Android Raw GNSS Measurement API, which allows access to raw GNSS data (pseudorange, pseudorange rate (Doppler), and accumulated delta range (carrier phase)) observed by the GNSS receiver embedded in smartphones, was released in 2017. This enabled sensor integration with smartphones's GNSS and paved the way for high-precision positioning with smartphones. To accelerate the development of high-precision positioning algorithms using smartphone GNSS observations, Google released a GNSS dataset in 2020 that was collected while driving with smartphones mounted on vehicles. Furthermore, the Smartphone Decimeter Challenge (SDC) has been held three times since 2021. As the name suggests, the SDC aims to estimate vehicle trajectories with decimeter-level positioning accuracy using raw GNSS observations from smartphones. In the third SDC, held from 2023 to 2024, decimeter-level positioning accuracy was finally achieved.

However, the positioning accuracy of smartphones remains limited to the decimeter level when in motion, whether in a vehicle or on foot. Resolving the integer ambiguity of carrier-phase observations is necessary to achieve centimeter-level accuracy, but this remains difficult for vehicle and pedestrian positioning on smartphones due to the constraints of smartphone GNSS antennas. Smartphone GNSS antennas are typically small, linearly polarized (LP) antennas that introduce issues in integer ambiguity resolution:

(1) Very low antenna gain increases GNSS observation noise, which degrades ambiguity resolution performance and prevents stable carrier-phase tracking during motion.
(2) LP antennas are sensitive to reflected multipath signals, which causes significant multipath errors in observations.
(3) Large phase center variation (PCV) introduces bias errors in carrier-phase observations depending on satellite elevation and azimuth angles.

Due to these issues, while ambiguity resolution has been partially successful in static, open-sky environments, achieving centimeter-level accuracy in kinematic positioning with smartphones has been difficult. Therefore, this paper proposes a method to boost the centimeter-level positioning accuracy of a smartphone's built-in GNSS receiver by developing a thin, zero-distance GNSS signal re-radiation system that can be attached to the smartphone. This system is constructed as a smartphone case, directly connecting a small, active Right-Hand Circularly Polarized (RHCP) helical antenna (used for standard GNSS positioning) to a thin, passive patch antenna for signal radiation. The amplified signal received by the active RHCP antenna is re-radiated directly above the smartphone's internal antenna. This system solves the following problems:

(1) Noise increase due to low gain: In the proposed system, the GNSS signal is received and amplified by a commercial active antenna before being re-input to the smartphone's antenna. This increases the GNSS signal strength observed by the smartphone, significantly reducing observation noise. It also enables stable reception of low-elevation satellite signals, reducing cycle slips and greatly improving carrier phase observation quality.
(2) Multipath problem: The proposed system uses an RHCP antenna to receive the signal, which greatly suppresses reflected multipath errors in the original signal. Although the smartphone's antenna still receives normal reflections, it primarily tracks the strong re-radiated signal, resulting in a significant reduction in multipath error.
(3) PCV problem: In the proposed system, all GNSS signals observed by the smartphone's antenna are transmitted from the phase center of the re-radiating antenna. As a result, phase variations due to the satellite's position do not occur, and the system effectively calculates the position based on the calibrated phase center of the RHCP antenna.

Through this smartphone- attachable GNSS signal booster device, centimeter-level positioning performance on smartphones can be greatly enhanced.

In this paper, we first designed and prototyped the smartphone-attachable GNSS signal re-radiation system. Signal re-radiation requires a power source to drive the active GNSS antenna. Therefore, we utilize the smartphone's wireless power supply system (Qi) and a receiving coil to power the active antenna contactlessly. Using a small helical antenna for signal reception and a thin passive patch antenna for signal transmission, we created a GNSS signal re-radiation system attachable to a smartphone. We used a Google Pixel 7 as the smartphone. Since the Pixel 7's GNSS supports L1 and L5 dual-frequency, the re-radiation system is also designed to transmit on the L1 and L5 dual frequencies.

To evaluate the system's performance, we first assessed the quality of smartphone observations in a static environment. Using the proposed system, a 5-7 dB-Hz increase in C/N0 was confirmed for all satellite signals. As a result of evaluating pseudorange observation accuracy using double-differenced pseudoranges, it was confirmed that the proposed system's pseudorange accuracy greatly improved with the increase in C/N0, leading to enhanced single-point positioning accuracy. Similar to pseudorange, it was also confirmed that the number of cycle slips decreased and carrier phase tracking became more stable. Furthermore, in carrier phase positioning, the integer ambiguity fix rate improved significantly, greatly enhancing centimeter-level position estimation performance.

Next, we evaluate the proposed system for pedestrian positioning. For the pedestrian test, a smartphone with the proposed re-radiation system, a standard smartphone, and a commercial GNSS antenna/receiver for reference measurement were mounted on a handheld pole, and we walked on an open-sky field. The results showed that the carrier phase ambiguity fix rate for the proposed system was significantly higher compared to the case without the system. In the actual presentation, we plan to show the evaluation of GNSS observation quality and the comparison of positioning accuracy for the proposed system in various environments.

In conclusion, this paper proposed a signal re-radiation system to boost the positioning accuracy of pedestrians using a smartphone's built-in GNSS. We plan to release the design information for the proposed system as open-source hardware after the presentation.

Previous Abstract Return to Session B6 Next Abstract