Estudo da distribuiÃÃo e autocorrelaÃÃo das observaÃÃes GPS

AUTOR(ES)
DATA DE PUBLICAÇÃO

2005

RESUMO

Relative GPS positioning is the most widely used technique current in geodesic surveying, and in fact one of the most precise. However, considerations concerning their functional and stochastic nature can result in overestimated precision, in other words, precision that looks better than reality. Among these considerations, two are objects of study in this research: the normality of the GPS observations, and the white noise process that involves the observable data. Therefore, it becomes necessary to study the occurrence of those factors in the GPS observations, with the objective of the development of better functional and stochastic models that estimate the true precision of GPS adjustments. With the residuals of each observable series obtained from the processing of a GPS baseline using the double difference method, statistical tests are described and applied that evaluate the hypothesis of normality of the distribution of each one of the observable series, as well as the hypothesis of the presence of a white noise process. In order to accomplishment these tests, the development of a processing GPS program was necessary that supplied each of the residual series. Applying the statistical tests, it was possible to verify that the GPS observations cannot always be considered as normally distributed, presenting different non-normal behaviors for each one of the observable series. Furthermore, the hypothesis of process white noise was rejected, showing the existence of autocorrelation in the data. Therefore, it appears to be a real necessity for implementating and bettering the current GPS models, to consider at least two factors in the processing of the GPS observations: the presence of non-normality and autocorrelations, thereby producing more realistic measures of GPS precision

ASSUNTO(S)

distribuiÃÃo das observaÃÃes gps distribution of the gps observations geodesia nonnormality posicionamento relativo nÃo-normalidade autocorrelaÃÃo gps autocorrelation gps relative positioning

Documentos Relacionados