基于综合灰关联序模型的残差门控循环神经网络位标器零部件选配

The drift performance of coordinator, one of the key factors that determined the tracking ability and guidance accuracy of the precision-guided weapon directly, might be improved by the selective assembly for coordinator parts. In view of the low success rate of the first selective assembly for coordinator parts, a CGRO model was proposed to perform the relevancy on assembly parameters affecting the selective assembly for coordinator parts, and key assembly parameters affecting the selective assembly for coordinator parts were obtained. The RNGRU model was established to realize the selective assembly for coordinator parts. Taking the selective assembly between the gyro rotor and drift-adjusting screw which affected the drift performance of the coordinator as an example, CGRO was used to analyze the assembly parameters of the gyro rotor, and the key assembly parameters that affected the selective assembly of the two were obtained. Based on RNGRU, the regression prediction of the mass of the drift-adjusting screw was realized. The experimental results show that the proposed method may achieve the selective assembly for coordinator parts, and the prediction accuracy is superior to the traditional GRU network.