提出了一种自适应无迹Kalman滤波(UKF)算法.针对UKF受初始值误差和动力学模型异常扰动误差影响的问题,将自适应估计原理引入到UKF算法,将动力学模型信息对导航解的贡献进行合理调整.计算结果表明,在GPS/INS松组合导航系统数据处理时,UKF算法略优于扩展Kalman滤波(EKF),自适应UKF算法优于自适应EKF算法,自适应UKF算法能够很好地抑制动力学模型误差对导航解的影响,进一步提高导航解的精度和可靠性.
A new adaptive unscented Kalman filter(UKF) algorithm is set up. In order to overcome the shortcomings of UKF, such as obvious influences from the values of initial parameters, the uncertainness of systemic noises and the influences of vehicle disturbances in movements, the adaptive estimation principle is applied for UKF. It is shown, by comparison and analysis that the UKF algorithm is better than extended Kalman filter(EKF) and the adaptive EKF is superior to UKF, and the adaptive UKF is superior to all algorithms in the application of GPS/INS integrated navigation systems.