← Tutorials

Tutorial Series · Part 21 of 27

Loose Coupling vs Tight Coupling

Compare fusing a receiver PVT solution against raw GNSS measurements in GNSS/INS integration.

By

  • gnss-ins
  • sensor-fusion
  • tight-coupling
Series Context

Navigation from First Principles

This article is part of an ordered course. Use the previous and next links to stay in sequence.

Progress Part 21 of 27
Previous Kalman Filter Intuition for Navigation Next GPS Outages: Tunnels, Cities and Jamming

GNSS/INS integration can happen at different levels. A loosely coupled system fuses the GNSS receiver’s position/velocity/time solution. A tightly coupled system fuses raw or preprocessed satellite measurements such as pseudorange and Doppler. Groves’ book is a useful broad reference for integrated navigation architectures.1

The problem

If a GNSS receiver cannot produce a full position fix, a loosely coupled filter may receive nothing useful. A tightly coupled filter can sometimes still use partial satellite measurements, because it does not require the receiver to output a complete PVT solution first.

Loose coupling

The GNSS receiver solves:

z^GNSS=[n^,e^,d^,v^n,v^e,v^d]T\hat{\mathbf{z}}_{GNSS}=[\hat{n},\hat{e},\hat{d},\hat{v}_n,\hat{v}_e,\hat{v}_d]^T

The navigation filter then uses this as a measurement:

z=Hx+v\mathbf{z}=\mathbf{H}\mathbf{x}+\mathbf{v}

This is simpler and modular.

Tight coupling

The navigation filter directly predicts measurements such as pseudorange:

Pi=∥r−si∥+cδtr+errorsP_i = \lVert \mathbf{r}-\mathbf{s}_i\rVert + c\delta t_r + \text{errors}

The filter update uses each satellite measurement individually. Navipedia’s GNSS observables and measurement-modelling pages are good supporting references for these raw measurement terms.23

Interactive demo: coupling architecture simulator

The demo below keeps the maths lightweight and focuses on the architectural distinction: loose coupling updates only when the receiver can produce PVT, while tight coupling can still use partial raw measurements in the three-satellite teaching case.

Interactive Demo

Loose vs Tight Coupling

Read the three views together: the route shows the estimate, the sky plot shows how many satellites survive, and the block diagram shows whether the filter receives a full receiver fix or only partial raw measurements.

Open sky Urban canyon Outage zone
Visible satellites
Satellites INS / IMU Receiver PVT ready Fusion Loose coupling: filter waits for receiver PVT.

Controls

Readouts

Receiver fixavailable
Measurement usefull update
End-of-route error0.0 m
What to noticeBoth modes track well while the receiver still has a fix.

What to watch

  • With four or more good satellites, both modes pull the estimate back toward truth.
  • When the receiver falls below its PVT threshold, loose coupling stops correcting and just coasts on INS.
  • Tight coupling can still extract some value from partial measurements when the receiver itself cannot declare a full fix.
  • The extra robustness comes with more measurement-modelling complexity.

Try this

Drive into the outage zone until the receiver no longer has enough satellites for PVT. In loose mode, the fused path simply follows inertial drift. Switch to tight mode and keep three satellites visible to see the partial correction survive.

Where this breaks

Tightly coupled integration is not automatically better in every implementation. It requires careful measurement modelling, satellite ephemerides, clocks, quality control and filter design. The right framing is trade-off, not slogan.

Footnotes

  1. Paul D. Groves, Principles of GNSS, Inertial, and Multisensor Integrated Navigation Systems, Second Edition, Artech House, 2013. https://uk.artechhouse.com/Principles-of-GNSS-Inertial-and-Multisensor-Integrated-Navigation-Systems-Second-Edition-P1557.aspx - Core textbook reference for navigation systems, GNSS, INS, multisensor integration, coordinate frames, dead reckoning, feature matching and environmental aiding. Accessed 2026-06-11. ↩

  2. ESA Navipedia, “GNSS Basic Observables”. https://gssc.esa.int/navipedia/index.php/GNSS_Basic_Observables - Reference for pseudorange and carrier-phase observables. Accessed 2026-06-11. ↩

  3. ESA Navipedia, “GNSS Measurements Modelling”. https://gssc.esa.int/navipedia/index.php/GNSS_Measurements_Modelling - Measurement equations including geometric range, clock terms, troposphere, ionosphere, multipath and noise. Accessed 2026-06-11. ↩