在惯性/天文组合导航系统中，天文设备输出信息中所包含的噪声易受外界环境影响而发生变化，使得惯性/天文器件级组合导航数据融合精度受到制约。针对这一问题提出了基于自适应滤波的惯性/天文器件级数据融合算法，该算法在对导航参数误差、惯性器件误差进行滤波估计的同时，对天文量测信息噪声方差阵进行实时推算，从而优化滤波器对天文量测信息的处理，提高导航精度。经仿真试验可知，使用自适应滤波算法进行数据融合，导航位置精度可提高30%，具有理论价值与工程意义。

In the system of INS/CNS integrated navigation, the noise of data from CNS always change along with the environments change, so that the result of navigation is not satisfied. In this thesis, we put forward automorphism INS/CNS component level integrated navigation algorithm, which can compute the parameter of observations noise as well as the errors of navigation and gyro. So that, the reformative algorithm effectively suppresses the various errors of INS. A series of tests show that automorphism INS/CNS component level integrated navigation algorithm can enhance the precision of orientation for 30%, thus the algorithm is significative for theory and application.