Development of two novel carrier phase-based methods for multiple reference station positioning
MetadataShow full item record
AbstractTwo methods for network-based carrier phase real-time kinematic positioning are proposed and evaluated in this thesis. These two methods are a correction-based, and a tightly coupled, approach. Novel algorithm enhancements are proposed for the correction-based approach, while the tightly coupled approach, which integrates reference station and user data in a single solution, is an innovative extension of a method previously developed. Each stage of the correction-based approach requires coherent information. The covariance function provides the stochastic basis of the estimation process and is used in each stage of this approach. A novel method for implementing an adaptive covariance function that is used subsequently is proposed. The adaptive qualities are shown to effectively track changing temporal and spatial error conditions, especially atmospheric conditions, throughout the data sets. The derivation of a least-squares prediction based approach, more specifically a least-squares collocation approach, is performed. This includes the value and variance-covariance of the estimated corrections. Further derivation shows the effect of these elements on the reduced rover measurements. This approach reduces the differential measurement errors and improves position accuracy relative to the single baseline approach. The tightly coupled approach is an extension of a multiple mobile user positioning approach, whereby inter-receiver position differences are connected to all reference stations and user(s) in the same estimation filter. This could also be considered an extension of the correction-based approach where the position of one or more of the reference stations is uncertain or unknown. This approach is also shown to improve position accuracy relative to the single baseline and correction-based approach. These two methods are compared and analysed. In general, both methods perform better than the single reference station approach however, the tightly coupled approach performs slightly better than the correction-based approach in terms of position accuracy, based on the data sets used for the evaluation of the two methods.
Bibliography: p. 197-203