Text
The 1st International Workshop on the Quality of geodesy observation andmonitoring system (QuGOMS'11) Proceddings of the 2001 IAG International Workshop, Munich, Germany, April 13-15,2011
This paper focuseson the modelingof data quality in processes which are in general not knowndetail or which are too complex to describe all influences on data quality. Asemerged during research, artificial neural networks (ANN) are capable formodeling data qualitparameters within processes with respect to their interconnections. Since multi-layer feed forward ANN are required forthis task a large number of examples, depending on the number ofquality parameters to be taken into account, is necessaryfor the suppervised learning of the ANN,respectively determining all parameters defining the net. Therefore the general usability of ANN was firstly evaluated for asimple geodetic application, the polar survey, where an unlimited number of learning examples could be generated easily. Aswill be shown, the quality parameters describing accuracy, availability, completenessor consistency and accuracy was tested as well. Standard deviations of new points can be determined using ANN wit-mm accurcy in all cases. To benchmark the usability of ANN forareal practical the complex process of mobile radio location and determination of driver trajectories on the digital road network based on these data, was used, The quality of calculated trajectories couiently from a number of relevan input parameters such ld be predicted sufficas antenna density and road density. The cross deviation asan impotant quality parameter for the trajectories could be predicted with an accuracy of better than 40 m.
B.20160708312 | 526.1 HAN t | Tersedia |
Tidak tersedia versi lain