惯导定位精度评价方法研究

    Study on the evaluation method of inertial navigation system positioning accuracy

    • 摘要: 随着惯导系统技术及产品的不断迭代,针对其定位精度的可靠的评定方法变得愈发重要。本文主要论述了95%位置圆概率误差统计方法(95%CEP)、真值均方根统计方法(TRMS)、95%置信度统计方法等三种算法的原理及特性,分析了三者之间的关系,并通过数模仿真及实验进行了验证,得出在设计好评定试验方案的情况下,95%CEP=1.73×TRMS,这一关系不因数据分布、样本点个数而改变;在数据分布服从正态分布时,95%置信度与95%CEP方法的统计值基本一致,与TRMS的关系在正态分布大样本时相对固定,小样本及非正态分布情况下变化较大;TRMS作为国外应用中的一个定位误差指标,具有评定方法简单、物理意义明确、统计实验参数少等特点,相对另外两种算法均具有优势。

       

      Abstract: With the continuous iteration of inertial navigation system technology and products, reliable evaluation methods for its positioning accuracy have become more and more important. This paper mainly discusses the principles and characteristics of three algorithms, including the 95% position circle probability error statistical method(95%CEP), the true root mean square statistical method(TRMS), and the 95% confidence statistical method, and analyzes the relationship between the three, and verifies it through the number imitation truth and experiments, and concludes that: 95%CEP=1.73×TRMS when the evaluation test scheme is designed, and this relationship does not change with the data distribution and the number of sample points; When the data distribution obeys the normal distribution, the 95% confidence level is basically consistent with the statistical value of the 95 P method, and the relationship with TRMS is relatively fixed when the normal distribution is large, and the change is greater in the case of small samples and non-normal distribution; As a positioning error index in foreign applications, TRMS has the characteristics of simple evaluation method, clear physical significance, and few statistical experimental parameters, which has advantages over the other two algorithms.