主要对扩展卡尔曼滤波（EKF）、无迹卡尔曼滤波（UKF）及改进无迹卡尔曼滤波（MAUKF）算法进行研究，研究了三种算法的基本原理和各自的特点。其中扩展卡尔曼滤波器是将卡尔曼滤波器局部线性化，其算法简单，计算量小，适用于弱非线性、高斯环境。无迹卡尔曼滤波器是用一系列确定样本来逼近状态的后验概率密度。改进无迹卡尔曼滤波算法在UKF的基础上引入衰减因子。-The thesis focuses on the extended Kalman filter (EKF), unscented Kalman filter (UKF) and the improvement of unscented Kalman filter (MAUKF) algorithm research, researched the basic principles of three algorithms and their characteristics. Extended Kalman filter (EKF) is the local linear Kalman filter, the algorithm and computation are simple in weakly nonlinear and Gaussian environment. Unscented Kalman filter (UKF) is determined using a series of samples to approximate the state posterior probability density. Improved UKF algorithm (MAUKF) is introduced attenuation factor based on the UKF.