A Confidence Assessment Method for Positioning Errors in Gravity-Aided Navigation

In the realm of gravity-aided navigation, the reliability of results generated by the TERCOM (Terrain Contour Matching) method can be influenced by the stochastic properties of gravity field distributions and measurement noise. In the scenario where the true position of the vehicle is unknown, it becomes challenging to assess the accuracy of the matching results, which can subsequently impact the corrective effect of gravity-aided navigation on INS (Inertial Navigation System). This study presents an innovative method to assess the confidence level of TERCOM outcomes. By conducting polynomial regression on historical TERCOM results, we create a fitting curve. The confidence level of the current result is then determined by evaluating its discrepancy from the fitting curve in terms of its percentile rank within historical discrepancies. Simulated experiments validate the utility of this approach, revealing a pronounced negative correlation between confidence levels and true positional errors, with a Spearman correlation coefficient of -0.8995. In the realm of precision evaluation for gravity matching outcomes, this quantitative methodology offers a concrete point of reference.