基于SVM的水下机器人舰位推算导航误差模型

Navigation Error Model Based on SVM for UUV's Dead Reckoning

本文为了提高UUV水下航行过程中舰位推算的导航精度,建立了基于支持向量机的导航误差模型。文章选取UUV的航向、相对大地的前向速度及左向速度、纵倾和横倾、导航推算时刻距离初始时刻的时间等六个与UUV导航误差密切相关的参数作为模型的输入,以舰位推算出的经度误差和纬度误差为模型输出,结果是验证该模型的有效性。结论是使导航的相对误差能够从48.69‰下降到1.49‰,得出误差模型可以较好地修正舰位推算中的经纬度误差。

To improve the process of UUV underwater ship reckoning navigation accuracy,the navigation error model is established based on support vector machine.This paper selects UUV heading,relative velocity of the earth and the left forward velocity,trim and heeling,navigation projected initial moment of time from the moment these six closely related parameter with the UUV navigation error are choosen as model input parameters to ship bit error calculate longitude and latitude error model output.It verifies the validity of the model.The relative error of the navigation can be dropped from 48.69 i（e） to 1.49 i（e）.The resulting model can be used to fix the error ship reckoning in longitude and latitude errors.