卡尔曼滤波在盾构姿态测量中的应用

Application of Kalman Filtering in Measurement of Shield Posture

三点法是一种传统的实时计算刚体三姿态角(滚动角、俯仰角和水平角)的方法。在盾构掘进过程中,观测点测量不同步会导致较大的姿态测量误差。采用卡尔曼滤波对观测点的坐标进行预测补偿来提高测量精度。建立系统运动的状态方程和量测方程,对系统噪声和量测噪声的协方差进行估计,通过递归运算对观测点坐标进行预测,蒙特卡洛仿真结果表明,这种处理方式优于传统三点法测量效果。仿真实验和测量实验结果都表明卡尔曼滤波算法既可以明显减小姿态角测量误差,又对随机系统噪声和量测噪声引起的姿态角误差具有抑制作用。

Three observation points method is a traditional real-time way for calculating attitude angles covariance of system noise matrix and measurement noise matrix are estimated to express the random noise of the system. Finally, coordinates of the observation points are predicted recursively. Monte Calro simulation results show that the performance of this method is better than those traditional methods. Both the simulation experiment and measurement experiment successfully demonstrated the proposed algorithm has the advantages of greatly decreasing the error for attitude angles, and inhibiting the effect of measurement error caused by random noise.