In GNSS/INS integration, how do loosely coupled and tightly coupled architectures differ?

• A. Loosely coupled uses GNSS position and velocity as filter measurements; tightly coupled fuses raw GNSS measurements (pseudorange, Doppler, carrier phase)
• B. Loosely coupled uses INS data only; tightly coupled uses GNSS data only
• C. Loosely coupled uses carrier phase only; tightly uses pseudorange only
• D. Loosely coupled uses raw GNSS measurements (pseudorange, Doppler, carrier phase); tightly coupled uses GNSS position and velocity

Answer: (A)

In GNSS/INS fusion, the way GNSS data is used within the navigation filter defines loosely versus tightly coupled integration. In a loosely coupled setup, the GNSS portion provides a standalone position and velocity (the GNSS solution), and these are treated as measurements that update the INS-based state through the filter. The INS and GNSS solution are fused at a higher level, with the GNSS solution serving as an external measurement update.

In a tightly coupled arrangement, the filter ingests the raw GNSS measurements themselves—pseudorange, Doppler, and carrier phase—directly, and fuses those with the INS information inside the same estimation process. This means the navigation solution is built from GNSS raw data and INS data together, even when not enough satellites are available to form a full GNSS solution externally.

This distinction makes the tightly coupled approach more robust to degraded GNSS visibility, since it can still estimate position using the raw measurements, whereas loosely coupled depends on a valid GNSS solution to provide updates. The stated correct description matches this: loosely coupled uses GNSS position and velocity as filter measurements; tightly coupled fuses raw GNSS measurements.