Abstract: In view of the detects existing in Kalman Filter and Extended Kalman Filter ( EKF) navigation，Unscented Kalman Filter( UKF) was introduced into GPS /DＲ integrated navigation system，and an improved window estimation algorithm was presented． An adaptive algorithm was used to regulate the window size，which could reduce the amount of calculation algorithm． The adaptive method was combined with UKF to track the system process noise in real time． The simulation results show that the proposed method has better precision，stability and adaptability，comparing with the traditional Unscented Kalman filter algorithm．

Key words: vehicle navigation, Unscented Kalman filter, GPS /DＲ, error estimation

摘要： 针对传统卡尔曼滤波和扩展卡尔曼滤波( EKF) 在导航中存在的缺点，在车辆GPS /DＲ 组合导航系统中引入了 Unscented Kalman filter( UKF) 算法，并提出一种改进的开窗估计算法，采用一种自适应算法来动态调整开窗大小，减 小了算法计算量。利用此方法对系统过程噪声进行实时跟踪，并与UKF 方法相结合。仿真试验表明: 此方法滤波 精度和稳定性都有了一定的提高，同时相比传统UKF 算法又具有更好的自适应性。

关键词: 车辆导航, 无迹卡尔曼滤波, 全球定位系统/航位推算系统, 误差估计