【摘要】 在运载体机动时,卡尔曼滤波器会导致捷联惯导系统初始对准时产生较大的机动误差。而H∞鲁棒滤波技术将噪声和不确定输入看作是只知道上界的参数,而不是统计特性服从某一分布的随机向量,从而提高了系统的鲁棒性。但由于H∞滤波一般不单独考虑系统不确定项的影响,从而得到的结果较为保守。因此,学者们又提出了离散时间不确定系统的鲁棒最小方差滤波技术。计算机仿真结果表明,用鲁棒H∞滤波器和鲁棒最小方差滤波器代替卡尔曼滤波器对SINS初始对准时的状态变量进行估计校正后,能有效降低SINS初始对准的机动误差,提高系统初始对准的精度。但鲁棒最小方差滤波器比鲁棒H∞滤波器滤波校正效果更好。更多还原

【Abstract】 When the carrier maneuvers, Kalman filter may result in large errors of SINS. However,H_∞ filter can decrease such errors since it regards uncertain parameters as randomly varying ones that have certain boundary instead of ones that obey some known statistical distribution. Nevertheless, H_∞ filter does not consider uncertain system items singly, which makes its filtering results conservative. Therefore, robust minimum variance filter has been proposed. The simulation results show that the errors existing in the initial alignment of SINS can be reduced effectively when the robust H_∞ filter and the robust minimum variance filter replace the Kalman filter. Thus, the accuracy of the whole system could be raised greatly. Furthermore, the performance of the robust minimum variance filter is better than that of the robust H_∞ filter.更多还原