微机械陀螺解耦控制技术的研究

Decoupling Control of Micro Mechanical Gyro

研究微机械陀螺优化控制问题。由于微机械陀螺具有很严重的耦合误差,使得其精度和性能大大减弱。现有微机械陀螺解耦方法大都集中在结构解耦,受工艺和外界环境影响严重,且稳定性较差。针对上述问题,为了从源头消除耦合,在结构解耦的基础上提出了两种控制解耦方案即传统解耦方法和BP神经网络解耦方法,首先用传统解耦方法在建立其精确数学模型的基础上找到合适的解耦矩阵,使得驱动的输入和检测的输出为单一方向的运动,来实现控制解耦。然后用传统解耦后的数据来训练BP神经网络,进行解耦仿真,从而根本上消除耦合误差。仿真结果表明,上述控制解耦方案能根本上有效消除耦合,提高微机械陀螺解耦控制的鲁棒稳定性。

The micro mechanical gyro has a serious coupling error, which makes its accuracy and performance greatly reduced. The existing micro mechanical gyro decoupling method mostly concentrates in the structure decoupling, which is severely affected by the process and the influence of the external environment, and has poor stability. In order to eliminate the coupling from the sources, two control decoupling schemes namely traditional decoupling method and BP neu- ral network deooupling method are put forward on the basis of the structural decoupling proposes. First, the right decou- piing matrix is obtained by using the traditional decoupling method based on the precise established mathematical model, so that the input and output are in single direction movement, and the decoupling control is realized. Then the traditional decoupled data are used to train the BP neural network, after decoupling simulation, the coupling error is fundamentally eliminated. The simulation results show that the decoupling control scheme can effectively eliminate the coupling fundamentally, and improve the robust stability of micro mechanical gyro deeoupling control.