摘要
In required navigation performance（RNP）, total system error（TSE） is estimated to provide a timely warning in the presence of an excessive error. In this paper, by analyzing the underlying formation mechanism, the TSE estimation is modeled as the estimation fusion of a fixed bias and a Gaussian random variable. To address the challenge of high computational load induced by the accurate numerical method, two efficient methods are proposed for real-time application, which are called the circle tangent ellipse method（CTEM） and the line tangent ellipse method（LTEM）,respectively. Compared with the accurate numerical method and the traditional scalar quantity summation method（SQSM）, the computational load and accuracy of these four methods are extensively analyzed. The theoretical and experimental results both show that the computing time of the LTEM is approximately equal to that of the SQSM, while it is only about 1/30 and 1/6 of that of the numerical method and the CTEM. Moreover, the estimation result of the LTEM is parallel with that of the numerical method, but is more accurate than those of the SQSM and the CTEM. It is illustrated that the LTEM is quite appropriate for real-time TSE estimation in RNP application.
In required navigation performance（RNP）, total system error（TSE） is estimated to provide a timely warning in the presence of an excessive error. In this paper, by analyzing the underlying formation mechanism, the TSE estimation is modeled as the estimation fusion of a fixed bias and a Gaussian random variable. To address the challenge of high computational load induced by the accurate numerical method, two efficient methods are proposed for real-time application, which are called the circle tangent ellipse method（CTEM） and the line tangent ellipse method（LTEM）,respectively. Compared with the accurate numerical method and the traditional scalar quantity summation method（SQSM）, the computational load and accuracy of these four methods are extensively analyzed. The theoretical and experimental results both show that the computing time of the LTEM is approximately equal to that of the SQSM, while it is only about 1/30 and 1/6 of that of the numerical method and the CTEM. Moreover, the estimation result of the LTEM is parallel with that of the numerical method, but is more accurate than those of the SQSM and the CTEM. It is illustrated that the LTEM is quite appropriate for real-time TSE estimation in RNP application.
基金
supported by the National Basic Research Program of China (No. 2010CB731805)
the Foundation for Innovative Research Groups of the National Natural Science Foundation of China (No. 60921001)
the Special Fund for Basic Research on Scientific Instruments of China (No. 2011YQ04008301)
