A new MEMS gyroscope drift suppression method for the low‐cost untwisting spin platform

The microelectromechanical system (MEMS) gyroscope provides a new method to design a low-cost untwisting spin platform to be used in a single-axis, stable strap-down inertial navigation system. However, the MEMS gyroscope's drift reduces the effectiveness of the closed-loop feedback control. Thus, a new method of drift suppression is proposed in this paper based on phase space reconstruction in order to improve the platform's performance. The feasibility of the MEMS gyroscope's drift suppression is analyzed using linear decomposition based on phase space. The system drift is estimated by phase space reconstruction. The optimal embedding dimension is found through a grid search. The number of dimensions for dimension reduction analysis is selected according to the minimum eigenvalue. The mapping from the high-dimensional phase space onto the low-dimensional phase space is obtained by minimizing the variance. A Kalman filter is used to compensate the residual sequence further. The proposed method is applied to an untwisting spin platform based on the MEMS gyroscope L3G4200D. The experimental results show that it can reduce the platform drift rate effectively.