Traditional artificial neural networks (ANN) such as back-propagation neural networks (BPNN) provide good predictions of length-of-day (LOD). However, the determination of network topology is difficult and time ...Traditional artificial neural networks (ANN) such as back-propagation neural networks (BPNN) provide good predictions of length-of-day (LOD). However, the determination of network topology is difficult and time consuming. Therefore, we propose a new type of neural network, extreme learning machine (ELM), to improve the efficiency of LOD predictions. Earth orientation parameters (EOP) C04 time-series provides daily values from International Earth Rotation and Reference Systems Service (IERS), which serves as our database. First, the known predictable effects that can be described by functional models-such as the effects of solid earth, ocean tides, or seasonal atmospheric variations--are removed a priori from the C04 time-series. Only the residuals after the subtraction of a priori model from the observed LOD data (i.e., the irregular and quasi-periodic variations) are employed for training and predictions. The predicted LOD is the sum of a prior extrapolation model and the ELM predictions of the residuals. Different input patterns are discussed and compared to optimize the network solution. The prediction results are analyzed and compared with those obtained by other machine learning-based prediction methods, including BPNN, generalization regression neural networks (GRNN), and adaptive network-based fuzzy inference systems (ANFIS). It is shown that while achieving similar prediction accuracy, the developed method uses much less training time than other methods. Furthermore, to conduct a direct comparison with the existing prediction tech- niques, the mean-absolute-error (MAE) from the proposed method is compared with that from the EOP prediction comparison campaign (EOP PCC). The results indicate that the accuracy of the proposed method is comparable with that of the former techniques. The implementation of the proposed method is simple.展开更多
We elaborate an error budget for the long-term accuracy of IGS(International Global Navigation Satellite System Service) polar motion estimates, concluding that it is probably about 25-30 μas(1-sigma)overall, alt...We elaborate an error budget for the long-term accuracy of IGS(International Global Navigation Satellite System Service) polar motion estimates, concluding that it is probably about 25-30 μas(1-sigma)overall, although it is not possible to quantify possible contributions(mainly annual) that might transfer directly from aliases of subdaily rotational tide errors. The leading sources are biases arising from the need to align daily, observed terrestrial frames, within which the pole coordinates are expressed and which are continuously deforming, to the secular, linear international reference frame. Such biases are largest over spans longer than about a year. Thanks to the very large number of IGS tracking stations, the formal covariance errors are much smaller,around 5 to 10 μas. Large networks also permit the systematic frame-related errors to be more effectively minimized but not eliminated. A number of periodic errors probably also influence polar motion results, mainly at annual, GPS(Global Positioning System) draconitic, and fortnightly periods, but their impact on the overall error budget is unlikely to be significant except possibly for annual tidal aliases. Nevertheless, caution should be exercised in interpreting geophysical excitations near any of the suspect periods.展开更多
基金supported by the West Light Foundation of the Chinese Academy of Sciences
文摘Traditional artificial neural networks (ANN) such as back-propagation neural networks (BPNN) provide good predictions of length-of-day (LOD). However, the determination of network topology is difficult and time consuming. Therefore, we propose a new type of neural network, extreme learning machine (ELM), to improve the efficiency of LOD predictions. Earth orientation parameters (EOP) C04 time-series provides daily values from International Earth Rotation and Reference Systems Service (IERS), which serves as our database. First, the known predictable effects that can be described by functional models-such as the effects of solid earth, ocean tides, or seasonal atmospheric variations--are removed a priori from the C04 time-series. Only the residuals after the subtraction of a priori model from the observed LOD data (i.e., the irregular and quasi-periodic variations) are employed for training and predictions. The predicted LOD is the sum of a prior extrapolation model and the ELM predictions of the residuals. Different input patterns are discussed and compared to optimize the network solution. The prediction results are analyzed and compared with those obtained by other machine learning-based prediction methods, including BPNN, generalization regression neural networks (GRNN), and adaptive network-based fuzzy inference systems (ANFIS). It is shown that while achieving similar prediction accuracy, the developed method uses much less training time than other methods. Furthermore, to conduct a direct comparison with the existing prediction tech- niques, the mean-absolute-error (MAE) from the proposed method is compared with that from the EOP prediction comparison campaign (EOP PCC). The results indicate that the accuracy of the proposed method is comparable with that of the former techniques. The implementation of the proposed method is simple.
文摘We elaborate an error budget for the long-term accuracy of IGS(International Global Navigation Satellite System Service) polar motion estimates, concluding that it is probably about 25-30 μas(1-sigma)overall, although it is not possible to quantify possible contributions(mainly annual) that might transfer directly from aliases of subdaily rotational tide errors. The leading sources are biases arising from the need to align daily, observed terrestrial frames, within which the pole coordinates are expressed and which are continuously deforming, to the secular, linear international reference frame. Such biases are largest over spans longer than about a year. Thanks to the very large number of IGS tracking stations, the formal covariance errors are much smaller,around 5 to 10 μas. Large networks also permit the systematic frame-related errors to be more effectively minimized but not eliminated. A number of periodic errors probably also influence polar motion results, mainly at annual, GPS(Global Positioning System) draconitic, and fortnightly periods, but their impact on the overall error budget is unlikely to be significant except possibly for annual tidal aliases. Nevertheless, caution should be exercised in interpreting geophysical excitations near any of the suspect periods.