A look at tethered satellite identification using ridge-type estimation methods

摘要

Recent studies of ill-conditioned orbit determination problems have demonstrated that ridge-type estimation methods have the capability to provide increased solution accuracy over classical estimation techniques. These advanced estimation methods add a small amount of bias to the solution in order to decrease the total variance of the estimates. Even though the addition of bias to the solution results in a slightly poorer data fit, the overall solution accuracy can be improved in certain types of ill-conditioned problems. One application in which the usefulness of ridge-type estimation methods has been demonstrated is the problem of quick-look orbit determination, i.e., the problem of accurately determining the trajectory of a spacecraft using only a short arc of observational data.One similar type of orbit determination problem which has received renewed interest lately is that of the identification of a satellite to be part of a tethered satellite system without prior knowledge that it is so. A batch filter has recently been developed which has the capability of identifying a satellite to be tethered, however the accuracy of this filter is highly dependent on the amount of observational data being processed. The performance of this filter has indicated that the shorter the data arc, the more ill-conditioned the problem, thereby resulting in inaccurate solutions.This paper applies ridge-type estimation methods to the problem of identifying a tethered satellite using short arcs of observational data. Performance of the ridge-type filter is evaluated for cases of differing tether lengths, tether orientations, levels of observation error, and observation span. Results will be compared to those obtained from classical minimum variance methods with special attention being given to the speed at which the identification of a tethered satellite can take place.