论文部分内容阅读
为了提高激光陀螺的性能,有效地补偿激光陀螺的随机误差,提出了一种灰色稀疏极端学习机的预测新模型。为了克服极端学习机(ELM)训练样本缺乏稀疏性的不足,根据矩阵求逆引理实现了冗余样本递推剔除,提出了一种稀疏极端学习机;利用灰色预测模型对原始数据进行初步预测,将实测值与预测值生成残差序列,选取适当长度残差序列作为训练样本,剩余数据序列作为测试样本,输入稀疏极端学习机进行残差回归预测,将预测的残差值与灰色预测得到的数据结合生成最终的预测结果。将灰色稀疏极端学习机预测模型应用于某型激光陀螺随机误差系数预测实验中,结果表明:该模型能够取得比其他3种预测模型更加精确的结果。
In order to improve the performance of the laser gyro and effectively compensate the random error of the laser gyro, a new prediction model of gray sparse extreme learning machine is proposed. In order to overcome the lack of sparsity of ELM training samples, a recursive elimination of redundant samples based on matrix inversion lemma was proposed and a sparse extreme learning machine was proposed. The gray prediction model was used to predict the original data , The residuals are generated from the measured values and the predicted values, the residuals of the appropriate length are selected as the training samples, and the remaining data sequences are used as the test samples, which are input to the sparse extreme learning machine for residual regression prediction. The predicted residuals and the gray prediction The data is combined to generate the final prediction. The gray sparse extreme learning machine prediction model is applied to the random error coefficient prediction experiment of a laser gyro. The results show that the model can obtain more accurate results than the other three prediction models.