【摘要】 针对INS/GPS/CNS组合导航系统的特点 ,采用了基于H∞ 滤波的神经网络代替传统卡尔曼滤波器 ,利用INS/GPS、INS/CNS两个局部滤波器的输出值 ,通过联合滤波器进行线性最优估计 ,且进行了计算机仿真。仿真结果表明 ,该算法在满足系统精度和容错性要求的基础上 ,有效地改善了系统的自适应性、实时性、鲁棒性等性能 ,是组合导航系统理想的结构模型。更多还原

【Abstract】 According to the characteristic of INS/GPS/CNS Integrated Navigation System, the paper constructed the filter with neural networks theory, and did linear optimal estimation with two local filter’s outputs. Simulation results show that the system improves not only adaptive capacity and real-time, but also the performance of robustness.更多还原