改进的UKF在GP S/INS组合导航中的应用

摘要

在GPS/INS组合导航中，传统UKF( Unscented Kalman Filter)计算量大，无法满足实时性要求。而且当动力学模型受到异常扰动误差影响时，其精度与稳定性易受到影响。针对以上问题，利用最小偏度单形采样策略降低UKF计算量以提高精度；通过自适应调整过程噪声以降低动态异常扰动误差对UKF精度与稳定性的影响。由此提出了一种改进UKF算法，用于GPS/INS组合导航。仿真实验结果表明，改进UKF算法用于GPS/INS组合导航的精度要优于UKF算法。%In the GPS/INS integrated navigation, the traditional unscented Kalman filter ( UKF) is unable to meet the real-time ability because of the large amount of calculation. What’ s more, the precision and stability of the tra-ditional unscented Kalman filter algorithm are susceptible to the effects of dynamic disturbance. To solve the prob-lem, the strategy of minimal skew simplex sampling can reduce the calculated amount and improve the accuracy. Adjusting the process noise adaptively can reduce the influence of the abnormal dynamic disturbance error on the accuracy and stability of UKF. An improved UKF was got from incorporating the two improved methods. The simu-lation experiment data analysis show that the precision of improved UKF integrated navigation algorithm is better than that of the UKF significantly.