一种提高捷联惯性/里程计组合导航系统定位精度的后处理方法

测绘用捷联惯性/里程计组合导航系统多采用离线后处理技术.此类系统利用里程计位移微分获得的速度作为观测量,采用速度匹配,并通过待测路径中预置的Mark点校正航位推算的位置信息.本文建立了基于速度匹配的16维Kalman滤波模型,对全程采样数据进行正反向导航和滤波处理,以估计惯性器件和里程计的误差;在补偿相关误差后,再次进行正向导航和滤波解算,以获得更精确的姿态矩阵.随后,根据相邻两个Mark点之间的姿态信息进行航位推算,并更新里程计刻度系数、里程计与惯导间的姿态误差矩阵;若此Mark点间的位置误差过大,则重新进行航位推算以减小位置误差.结果表明,与传统正向滤波相比,采用该方法后系统的最大位置误差由1.72m降低到0.08m,定位精度提高了95％以上.

An Post-processing Method of Enhancing Position Accuracy of Strapdown Inertial/Odometer System

The offline post-processing technology is commonly used in strapdown Inertial/Odometer integrated navigation systems. Based on velocity matching filter, this kind of system employs velocity by differentiating odometer''s displacement, adopts velocity matching and corrects the dead reckoning position by utilizing pre-set Mark point in the route to be measured. In the first stage of this method, a 16-dimension velocity-based filter model is established and a forward and reverse filter algorithm is applied, which are used for processing the whole offline sample data, thereby estimating inertial sensors'' errors and the odometer''s errors. After compensating relevant errors, the forward navigation process from initial point is recalculated. This process, as the second stage of this method, is perceived as an efficient way to approximate the attitude matrix. In the third stage, according to attitude information of adjacent Mark point, the Inertial/Odometer integrated navigation system is reckoning carriers'' position, updating odometer''s scale factor, and correcting installation errors between odometer and inertial navigation system. When position error reaches the threshold, re-reckoning process should be executed to reduce position error. By utilizing this method, the maximum position error is minimized from 1.72m to 0.08m, and the position accuracy is increased by more than 95%.