A SVM Based GNSS Performance Assessment with Reliable Vulnerability Degree Model

Yuankang Chen ； Xingqun Zhan ； Jiaxun Tu

《Journal of Aeronautics, Astronautics and Aviation》 50卷3期 Vol.50 No.3 (2018/09) Pp. 301-314

https://doi.org/10.6125/JoAAA.201809_50(3).07

摘要

In recent years, an increasing number of GNSS vulnerability assessment methods have been formulated in order to inform the users of the performance of GNSS service. However the reliabilities of these assessment methods are hard to be guaranteed due to the difficulty to verify the accuracies of assessment results. A novel model of vulnerability degree is proposed here in order to make the reliabilities of these methods able to be validated. In this model, the vulnerability degree is determined by positioning error which reflects the GNSS service quality directly. With massive collected GNSS data, the accuracy of the typical method using the theory of Dempster-Shafer evidence reasoning is tested based on the new model. For further improvement, Support vector machine (SVM), choosing 9 essential metrics, is introduced to improve the accuracy of GNSS vulnerability assessment. As the validation result shows, the accuracy of the new method using SVM is better than the classical method using the theory of Dempster-Shafer evidence reasoning based on the new model of vulnerability degree, just at the cost of a little increase of resource consumption.

關鍵字

GNSS vulnerability ； vulnerability degree ； SVM ； Evidence reasoning