Vlbi, Slr, Gps Data Processing And Coordinate Conversion A Number Of Issues

This dissertation presents th7 theory and method research on the combinationadustment of VLBI, SLR and GPS data. Some problems on datum transformationare also studied. First, the basic princiPle of the three spatial geodehc measuretecboques is introduced and the proPenies of the Cwr establishCd by individualtecndques are shOwn. As fOr daum transformation, we present a-coordinatetransfoMon model of three-scale paramters based on the hyPothesis that thescale space is isotropic and its speci8l instances, namely, the tWo-scale and non-scaleparamter coordinate transformation modls. A united hyPothesis-testing modl ofscale parameter is given and thC corresPondng ndshc of testing is constructed.Considering the local error accumulation and distonion, a combined method ofsinilledty transformation and regressive aPPrsimating is present and the precise ofcoordinate system is linProved after the transformallon using the reconunendedmethod.General1y,. two methods of combination have been proPosed and used in realdata processing. One is tO process all raw data setS simultaleously, and the otherprocesses the quasi-observables which are the cOmPressed form of each data session.It is proved that the tWo methods of combination are actua1ly equivalent to eachother in the dissertaion. According to the tyPes of observation value, we classify theschemes of combination adjustment into four kinds whose advanages anddisadvantages are analyzed. The suitable schCme is proPosed accordng to the realsituation of the three observation techniques. Some prob1ems suCh as the unificationof epoch and coordinate datUm and the statishc testing of systematic parameters arealso present.As for the stochashc model of combination adjustment, we discuss the errorsespecially the systematical error of stOChastic mode1 iIl this dissertation. We presenta review of the methods for estimating the vedance cOmPonents, which can bedishnguished as functinnal medels, stochashc models, eshmation aPProaChs, andngorous or simPlified procedures within the context of the LS adjustment develOpedso far In order tO resist the effeCs of OUiling Observation, a robust HelInert'svanance comPonents eshmation is introduced. Ths estimation strategy can not ontyresist the effects of ouhying observation on the eshmates of the model paramters,but also give a reasonable variance comPonens, naInel, a reasonable weightS fOrdiffernt observation tyPes. By requiring the residuals of combinahon solution to bewhite Gaussian as a samP1e of random noise, on the other words, no systematicinformation tO be detected in them, we can assess the combination soluhon. T'hehason x'' higher-Order moments and aUtoCorrelation coefficients of residuals areintrOudced and used tO measure qUantitatively the diWancies between theresiduals and white Gaussha nose in this disedon. Finally, some sboulativecaculations on the combination adustInent of VLBL sLR and GPS data are madeand some usend comclusions are obtained.