Gravity model evaluation for precise terrestrial inertial navigation- A system accuracy approach

Abstract

Abstract : Inertial navigation has evolved to the point that the traditional gravity model is a principle error source in precise systems. Since future systems may require a more accurate gravity model, a statistical performance evaluation is needed to compare candidate models on realistic scenarios. The present methods of model evaluation are Monte Carlo, which is computationally costly, and linear state space covariance analysis, which has troublesome trajectory and statistical model restrictions. This research developed a new statistical analysis technique based on a double integral expression of the navigation system error covariance. Numerical algorithms were formulated to approximate the covariance, and the accuracy and efficiency of these algorithms were compared using a simple Schuler Loop example analysis. (Author)