GNSS in Open Sky vs Urban Environments: A Performance Comparison

Introduction

"GNSS works beautifully in open sky - but real-world environments are far more complex."

GNSS positioning performance is not uniform. The environment surrounding the receiver antenna fundamentally determines what accuracy, availability, and reliability are achievable. While datasheets may quote impressive positional figures, those numbers almost always assume open-sky conditions - a scenario that represents only a fraction of real-world deployments. Engineers designing systems for urban delivery vehicles, precision agriculture at field edges, or surveying near structures must understand how and why performance degrades as the environment changes.

Open Sky Conditions

An open-sky environment is defined as one where no significant obstructions block the satellite signals above a defined elevation cut-off angle, typically 10–15 degrees above the horizon. In these conditions, the receiver has access to signals from a large number of satellites distributed across the sky in multiple directions and at varying elevation angles. The geometry is favourable, Dilution of Precision (DOP) values remain low, and signals arrive via direct line-of-sight (LOS) paths.

Under open-sky conditions, standard GNSS code-phase positioning achieves horizontal accuracies of 1–3 metres, while carrier-phase techniques such as Real-Time Kinematic (RTK) routinely deliver sub-centimetre results. Signal quality is consistent, carrier tracking is stable, and the receiver can maintain a robust navigation solution with redundant satellite measurements.

Urban Canyon Behaviour

An urban canyon describes an environment where tall buildings flank both sides of a street or corridor, severely restricting the visible sky to a narrow strip directly overhead. Research across multiple cities has confirmed that this is the most challenging common GNSS environment. The three dominant degradation mechanisms in urban canyons are:

• Signal Masking: Buildings obstruct large portions of the sky, reducing the number of satellites in view. Studies have found that in typical urban canyons, the number of visible satellites can fall below four - the minimum required for a 3D position fix - for extended periods.
• Multipath Interference: Building facades reflect GNSS signals. The receiver receives both the direct signal and delayed reflected copies, which corrupt pseudorange and carrier-phase measurements without triggering obvious quality flags.
• Non-Line-of-Sight (NLOS) Reception: In the most severe cases, the direct signal is completely blocked and only a reflected path reaches the antenna. The receiver may still track this signal and use it in the navigation solution, producing large, systematic errors.

The combined effect is dramatic. Studies in Hong Kong urban canyons have demonstrated horizontal RMSE values exceeding 13 metres in the worst-affected areas, compared to sub-metre performance in open-sky tests at the same epoch. These are not outlier conditions - they represent everyday operating environments for fleet management, personal navigation, and autonomous vehicle systems.

Satellite Visibility Differences

The contrast in satellite visibility between open-sky and urban environments is stark. In open sky with multi-constellation GNSS (GPS + GLONASS + Galileo + BeiDou), a receiver may track 30 or more satellites simultaneously. In a deep urban canyon, that number can drop to fewer than six, with no guarantee that the visible subset provides adequate geometric spread.

Even where multi-constellation receivers help by increasing the pool of potentially visible satellites, NLOS and multipath signals from additional constellations can worsen the solution if they are incorporated without proper quality checking. More signals are only beneficial when they are valid LOS measurements.

Accuracy Comparison

Why Environment Defines Accuracy

The fundamental point for GNSS engineers is that the environment is not a secondary consideration - it is the primary determinant of achievable performance. A receiver with excellent hardware, multi-constellation capability, and advanced algorithms will still produce metre-level or worse errors in a deep urban canyon if no additional measures are taken. This is not a receiver deficiency; it is a physical consequence of the propagation environment.

Systems deployed in mixed environments must be designed with this in mind. Position quality indicators, sensor fusion with IMU or odometry, map-matching algorithms, and architecture-level mitigation strategies are all necessary tools when urban performance matters. The open-sky specification on a datasheet is a starting point for understanding a receiver, not a guarantee of field performance.