Optimizing GNSS Time-Differenced Carrier Phase Measurements for High-Integrity Inertial Navigation + GNSS Sensor Fusion without Ambiguity Resolution

Typical inertial sensor integration uses a single GNSS receiver which has a positioning accuracy between 10m and 2m, or, for high-accuracy survey in ‘RTK’ mode, which provides cm-level precision by adding a 2
^nd
, stationary GNSS base station. Time-differenced carrier-phase (TDCP) displacement measurement is a powerful measure of the receiver displacement using consecutive GNSS measurements with up to mm-level precision, without the need for any external base station by differencing raw GNSS satellite signals between rover measurements. This paper it presents and compares two solutions to improve the reliability of the TDCP, which is required for high-precision INS/TDCP coupling. In stand-alone GNSS applications, the assumptions required to use TDCP can lead to displacement accuracy being highly degraded. For a high-integrity solution, TDCP measurement reliability and the precision of the calculated displacements must be optimized. Two methods are investigated here: 1. By considering the large number of raw satellite information epoch-to-epoch and applying robust regression, and 2. Using an estimation of the inertial sensor displacement over a short integration period to detect outliers. When coupled with an INS, this optimized TDCP displacement can be useful to contain the position drift when a precise GNSS (RTK, PPP) is not available. The displacement calculated from TDCP measurements must be highly precise and reliable, even at the expense of availability. To make these optimizations, a data driven method is used, taking measurements from, and making comparisons to, a robust database of real-world “truth” data collected in a variety of environments.