Integration of 5G and Motion Sensors for Vehicular Positioning: A Loosely-Coupled Approach
Autonomous vehicles (AVs) are poised to revolutionize the transportation industry by enhancing traffic efficiency and road safety. However, achieving optimal vehicular autonomy demands an uninterrupted and precise positioning solution, especially in deep urban environments. 5G mmWave holds immense p...
Gespeichert in:
Hauptverfasser: | , , , , |
---|---|
Format: | Artikel |
Sprache: | eng |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Autonomous vehicles (AVs) are poised to revolutionize the transportation
industry by enhancing traffic efficiency and road safety. However, achieving
optimal vehicular autonomy demands an uninterrupted and precise positioning
solution, especially in deep urban environments. 5G mmWave holds immense
potential to provide such a service due to its accurate range and angle
measurements. Yet, as mmWave signals are prone to signal blockage, severe
positioning errors will occur. Most of the 5G positioning literature relies on
constant motion models to bridge such 5G outages, which do not capture the true
dynamics of the vehicle. Few proposed methodologies rely on inertial
measurement units (IMUs) to bridge such gaps, where they predominantly use
tightly coupled (TC) integration schemes, introducing a nonlinear 5G
measurement model. Such approaches, which rely on Kalman filtering, necessitate
the linearization of the measurement model, leading to pronounced positioning
errors. In this paper, however, we propose a loosely coupled (LC) sensor fusion
scheme to integrate 5G, IMUs, and odometers to mitigate linearization errors.
Additionally, we propose a novel method to design the process covariance matrix
of the extended Kalman filter (EKF). Moreover, we propose enhancements to the
mechanization of the IMU data to enhance the standalone IMU solution. The
proposed methodologies were tested using a novel setup comprising 5G
measurements from Siradel's S_5G simulation tool and real IMU and odometer
measurements from an hour-long trajectory. The proposed method resulted in 14
cm of error for 95% of the time compared to 1 m provided by the traditional
constant velocity model approach. |
---|---|
DOI: | 10.48550/arxiv.2403.10872 |