Hardware and System Biases in Multi-Constellation Positioning

30 concepts16 papers1 proposal

In an ideal positioning system, calculating a location is a pure geometric problem. By measuring the time it takes for signals to travel from multiple known points (like satellites) to a receiver, we can draw intersecting spheres to pinpoint an exact location. To make this solution more robust, the obvious strategy is to add more signals—combining different satellite constellations (such as GPS, BeiDou, and QZSS) or augmenting them with Low Earth Orbit (LEO) satellites and terrestrial sensors.

However, in the real world, this geometric elegance is disrupted by a hidden variable: physical hardware. Before a signal ever travels through space, it passes through internal circuitry, filters, and cables within the transmitter. Once it arrives, it undergoes similar processing inside the receiver. Each of these steps introduces minute, uncalibrated timing delays. Because light travels at roughly 30 centimeters per nanosecond, a hardware delay of just a few nanoseconds translates directly into meters of positioning error.

When we mix signals from different constellations, frequencies, or hardware grades, these delays do not cancel out. Instead, they manifest as systematic biases that can severely degrade positioning accuracy if left unaddressed.

The Origin of the Bias: Satellite vs. Receiver Hardware

Systematic biases enter the positioning equation from both ends of the signal path.

On the transmission side, satellites suffer from internal hardware delays that vary by frequency and signal structure. As shown by Cheng Yang 0005 et al., these satellite-specific systematic errors—particularly in clock offsets—directly degrade the Signal-in-Space Range Error (SISRE) of precise positioning services, meaning that even advanced correction data cannot fully resolve inherent satellite-level inaccuracies. Furthermore, these biases can be signal-dependent. For instance, Xiaping Ma et al. highlight the impact of Satellite-Induced Code Bias (SICB) in the BeiDou (BDS) constellation, demonstrating that standard smoothing algorithms suffer when these satellite-level variations are ignored.

On the receiving side, the quality of the hardware dictates how cleanly these signals are processed. High-end geodetic receivers use sophisticated components to minimize and stabilize internal delays. In contrast, consumer-grade hardware introduces massive volatility. Zeyu Ma et al. found that low-cost receivers exhibit substantially larger and more variable biases than geodetic equipment. This is most acutely felt in smartphones; Xu Weng et al. demonstrate that raw GNSS pseudorange measurements from smartphones are plagued by hardware-induced biases that can reach up to 10 meters, severely bottlenecking standard positioning algorithms.

The Multi-System Challenge: Inter-System and Inter-Frequency Biases

The problem compounds when a receiver attempts a "multi-constellation" solution. If a receiver tracks both a GPS satellite and a BeiDou satellite, it cannot treat their timing data identically.

First, different constellations operate on different internal time scales and use distinct signal architectures. This mismatch introduces Inter-System Biases (ISBs). As explored by Qiuzhao Zhang et al. and Qian Liu et al., accurately estimating and isolating ISBs is critical for multi-GNSS fusion, especially when operating in challenging environments with limited satellite visibility.

The physical reason for these biases often traces back to how receivers handle time. As Houzhe Zhang et al. explain, different global navigation systems utilize differing time tag correction conventions. If a processing engine forces a single, unified receiver clock offset across all systems, these time-tag discrepancies warp the mathematical solution. Instead, algorithms must estimate independent receiver clock offsets for each constellation to absorb the inter-system friction.

Second, even within a single satellite system, utilizing multiple frequencies introduces Inter-Frequency Biases (IFBs) and observable-specific biases (OSBs). Because different frequencies travel through different physical analog paths inside the hardware, they experience distinct phase and code delays. Xin Li 0090 et al. note that as multi-frequency, multi-GNSS data grows more complex, the parallel estimation of these overlapping inter-frequency clock biases becomes computationally essential to maintain real-time positioning viability.

The Danger of Performance Inversion

Intuitively, adding more data should always improve an optimization problem. But in biased systems, the opposite can happen: a phenomenon known as performance inversion.

If you integrate auxiliary signals that contain uncalibrated hardware delays, the positioning engine treats those systematic delays as actual geometric distances. The optimization algorithm then warps the entire coordinate solution trying to accommodate the faulty data. Kan Wang et al. demonstrated this by augmenting standard GNSS with LEO satellite signals. If the internal hardware timing delays of the LEO satellites exceed a specific threshold and remain uncalibrated, the augmented system actually performs worse—yielding lower accuracy and slower convergence times—than a baseline solution that completely ignores the extra satellites.

Without Bias Calibration:
Standard GNSS Signals  ───┐
                          ├──► Optimization Engine ──► Warped Geometry ──► Lower Accuracy
Uncalibrated LEO Signals ───┘

With Bias Calibration:
Standard GNSS Signals  ───┐
                          ├──► Bias Estimation Filter ──► Clean Geometry ──► Enhanced Accuracy
Uncalibrated LEO Signals ───┘

When these uncalibrated system biases propagate through the positioning engine, they do not distribute evenly across all coordinates. Because of satellite geometry—where satellites are always distributed across the sky above the user but never below them—systematic biases heavily skew the vertical component of the estimated position. Yunhao Chen et al. confirmed this asymmetry globally, showing that unaddressed system biases in BeiDou-3 signals manifest prominently as a consistent offset in the upward (vertical) direction.

A Universal Localization Problem

This challenge is not unique to satellites; it is a fundamental property of any ranging hardware.

In terrestrial Ultra-Wideband (UWB) indoor positioning systems, which rely on Time-Difference-of-Arrival (TDoA), low-cost chips introduce similar systematic measurement biases. As Paolo Grasso et al. and Wenda Zhao et al. observe, even when a UWB system functions perfectly, inherent chip-level biases degrade localization accuracy unless dedicated debiasing filters are applied. The same pattern emerges underwater: Eduardo Alexander Iscar Ruland et al. identified a consistent hardware range offset in one-way travel time acoustic localization systems, proving that whether a signal is electromagnetic or acoustic, hardware introduces systematic friction.

Restoring Precision through Separation and Estimation

To unlock the true potential of multi-constellation positioning, modern algorithms must explicitly decouple hardware biases from environmental errors (like ionospheric delays) and geometric coordinates.

For example, in sparse reference networks used for high-precision real-time kinematics (PPP-RTK), ionospheric delay measurements are naturally coupled with receiver hardware biases. Sijie Lyu et al. resolved this by designing regional models that mathematically separate individual receiver biases from the ionospheric state. By isolating the hardware datum inconsistencies, they prevent receiver-specific errors from corrupting the shared network corrections.

Ultimately, whether by utilizing parallel state-space estimation filters to track clock biases in real time, accounting for signal structure differences in evolving constellations, or deploying neural networks to predict smartphone pseudorange errors directly, the core insight of the field remains the same: to achieve centimeter-level accuracy, a positioning system must map its own hardware just as precisely as it maps the open sky.

Key Foundational Work

Performance Evaluation of QZSS Augmenting GPS and BDS Single-Frequency Single-Epoch Positioning with Actual Data in Asia-Pacific Region (2018) — contributed 1 concept elements to this direction.
Low cost underwater acoustic localization (2017) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_satellite_specific_systematic_errors_58d68389,
  title = {satellite-specific systematic errors},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/satellite-specific-systematic-errors/rdr_58d683894784405bb90df5e37ffc7ea2},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_satellite_specific_systematic_errors_58d68389_v1,
  title = {satellite-specific systematic errors},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/satellite-specific-systematic-errors/rdr_58d683894784405bb90df5e37ffc7ea2/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake