Evaluation of the optimal design 'cosinor model' for enhancing the potential of robotic theodolite kinematic observations

摘要

The aim of the present work is to assess and to demonstrate the benefits of adopting optimal experimental design theory and techniques in order to enhance the potential of field data recorded using conventional geodetic instruments. More specifically, this research focuses on Robotic Total Stations (RTS) and in kinematic applications of geodetic positioning that exhibit a cyclic/periodic pattern of motion. The computational approach adopted follows from the principles of c- and D-optimal design criteria. Data processing involves computing the amplitude of motion in two ways; (a) using a sample of consecutively ordered data recordings and (b) using a sample respecting the optimal design criteria. Analysis, confirms the utility of the method resulting an improvement (i.e. a reduction) of the oscillation amplitude variance. This conclusion applies particularly at higher frequencies of oscillations (>1 Hz). This is important as at higher frequencies the performance of RTS deteriorates, and hence large variances occur.