新测量船惯导数据融合使用方式研究

Research on Inertial Reference Data Fusion for New Space Instrumentation Ship

新一代远洋航天测量船安装有分布式捷联惯导系统,因此具备对某个点位的姿态提供多个测量数据的能力。这不仅加强了数据的可靠性,同时也为融合使用数据提供了可能。分析了如何由多个点位的姿态独立测量数据生成同一位置的冗余姿态量,使用了基于最大似然估计的数据融合法对船体姿态数据进行融合。在方法实践过程中,发现固定的方差参数并不适用于长时间的测量过程,由此采用动态生成方差法对原方法进行了改进。最后通过海上试验数据对实测数据、融合数据和经改进后融合数据进行了误差统计与分析,分别从正常数据和异常数据两方面对方法进行了验证。

New space instrumentation ship with distributed strapdown inertial navigation system has the capability that provides several attitude data for one position.This capability not only reinforces the reliability of data,but also provides possibility for data fusion.The process which uses multipoint attitude data to generate singlepoint redundancy attitude data is analyzed.The ship attitude data are fused by a data fusion method based on maximum likelihood estimation.The fixed variance parameters don＇t adapt with long time measuring process.The fusion method is improved by dynamic calculating variance based on this characteristic.Finally,the random error of measured data,fusion data and fusion data generated by improved method are input to statistical analysis.The improved fusion method is verified by normal data and abnormal data.