Design and assessment of horizontal survey networks

Abstract

The research reported in this thesis covers the investigation and implementation of techniques to aid in the design and assessment of horizontal survey networks. To date, much of the effort in the design of survey networks has been towards the optimization of the precision of a network, as represented in the covariance matrix of the coordinate variates. While a network may be designed to be very precise, it may not be accurate due to the presence of non-stochastic errors in the observations. Part of this research covers the theory and application of reliability analysis to give estimates for the sensitivity of networks to the presence of non-stochastic errors in the observations. These methods then enable the design of networks to be both precise and reliable. Also presented in this thesis is an assessment scheme for the detection and localization of outliers in survey network adjustment results. After the screening of adjustment results, the recovery of any miscaling of the a priori variances of the input observations may be obtained through application of variance component estimation techniques. To aid in the detection of movements in repeatedly measured procedure networks, a detailed design and assessment is presented to aid in the detection of movements between measurement epochs that are in the order of the standard deviation of the input observations.