Smoothing procedures for inertial survey systems of local level type

Abstract

The research that is reported in this thesis investigates the suitability and accuracy of smoothing procedures for a Local Level Inertial Survey System (LLIS). The study concentrates on the comparison of optimal and empirical smoothing procedures, taking Litton's Autosurveyor as a typical example of the latter. The analysis is carried out in four stages. In the first stage, a set of differential equations describing the navigation system error behavior is derived. The second stage discusses the implementation of Kalman filtering. The third stage considers the implementation of the basic equations of optimal and empirical smoothing. In the fourth stage, accuracy figures of optimal smoothing and empirical smoothing for different traverse configurations are derived using digital simulations of a LLIS. The results of this research show that empirical smoothing is an attractive method of processing filtered data. A comparison of accuracy figures between optimal and empirical smoothing algorithms derived from simulated data has shown no significant differences. The simple equations of empirical smoothing eliminate the need of storing a large amount of data, avoid the lengthy computations needed for optimal smoothing, and can be implemented on a small computer.