改进抗差Kalman滤波的INS/GPS紧组合导航算法

INS/GPS Tightly Coupled Navigation Based on Improved Robust Kalman Filter

现有的抗差卡尔曼滤波基于对位置信息的检验给出等价权,会导致当GPS可观测卫星在4颗以下时,基于抗差卡尔曼滤波算法的INS/GPS紧组合导航算法退化为纯惯性导航,针对这一问题,提出了改进的抗差卡尔曼滤波算法.首先介绍了INS/GPS组合导航模式,并且建立了紧组合导航模式下的状态空间方程;然后分析了GPS可观测卫星在4颗以下时,抗差卡尔曼滤波导航精度差的根源,在此基础上提出了改进的抗差卡尔曼滤波算法,改进的算法基于原始测量信息给出等价权,因此无需可观测卫星在4颗以上;实验结果表明,改进的抗差卡尔曼滤波克服了抗差卡尔曼滤波的缺陷,在导航精度上有很大提高.

Exsisting Robust Kalman Filter establishes equivalent right based on positional information,which means when the number of observable satellites is less than four,Robust Kalman Filter degrades to pure inertial navigation.Aiming at solving the problem,this paper proposes an improved Robust Kalman Filter algorithm.INS/GPS coupled navigation model is introduced,and the state space equation of tightly coupled navigation system is built.When the number of observable satellites is below four,the root reason of poor navigation accuracy of Robust Kalman Filter is analyzed to put forword an improved Robust Kalman Filter.Equivalent right in the new algorithm is calculated based on new measurement,so that this algorithm has no requirement of more than four observable satellites.Experiment is designed to verify that the improved Robust Kalman Filter can overcome the shortcoming and has higher navigation accuracy.