GNSS/INS integration is a form of sensor integration or sensor fusion, which involves combining the outputs of different sensor systems to obtain a better estimate of what they are sensing.
A GNSS receiver is a position sensor. It may use velocity estimates to reduce filter lags, but its primary output is the position of its antenna relative to an earth-centered coordinate system. GNSS position errors will depend on the availability and geometric distribution of GNSS satellites it can track, and other error sources described in Chapter 5. The resulting RMS position errors will be bounded, except for those times when there are not enough satellite signals available for a position solution.
An INS uses acceleration and attitude (or attitude rate) sensors, but its primary output as a sensor system is also position—the position of its ISA relative to an earth-centered coordinate system. INS position errors depend on the quality of its inertial sensors and earth models, described in Chapter 9. Although their short-term position errors are very smooth, RMS position errors are not bounded. They do tend to grow over time, and without bound.
This chapter is about practical methods for combining GNSS and INS outputs to improve overall system performance metrics, including