一种智能的惯导陀螺漂移误差消除算法

An Intelligent Algorithm of Eliminating INS Gyroscope Drift Error

给出了一套智能的消除惯导陀螺漂移误差算法,即改进的BP(Back Propagation)神经网络模态识别算法。可根据即时的飞行区域、气压计和雷达高度表测量的地形高程数据等,在机载数字地图上实时地进行在线模态识别,找出最合适的匹配区,并对惯性导航系统指示的位置信息进行修正。该算法具有识别精度高、识别速度快和识别准确率高等特点,可用于各类有人驾驶飞行器和无人自主控制的飞行器的惯导误差修正。

This paper presents an intelligent pattern recognition algorithm （IPRA） with the updated BP neural network on eliminating the INS （Inertial Navigation System） gyroscope drift error. The algorithm incorporates the effects of flight region and measured terrain height data by radar and barometer. Based on this algorithm, the appropriate match region was gotten by recognition of fiducial digital map in real time online. Therefore, the INS error can be updated by IPRA for both manned aircraft and unmanned flight vehicles. The algorithm possess rapid recognition velocity, high recognition precision and correct probability of recognition.