The Filtering Technology of Virtual Gyroscope Based on Taylor Model in Low Dynamic State

In view of the large output noise and low precision of the micro-electro-mechanical system (MEMS) gyroscope, the virtual gyroscope technology is used to fuse the data to improve its output precision. However, the traditional virtual gyroscope technology has poor dynamic compensation effect. Therefore, a virtual gyroscope technology with high-precision dynamic filtering is proposed. First of all, the high-accuracy model of the real dynamic angular rate based on Taylor formula is established, and the random error model of the MEMS gyroscope array is improved based on the real angular rate modeling results. The Kalman filter is designed for dynamic data fusion to improve the dynamic compensation precision of the virtual gyroscope. The test results show that the $1\sigma $ standard deviation of the estimation error of the proposed method is two times lower than the $1\sigma $ standard deviation of the direct KF model’s dynamic estimation error and it is nearly eight times lower than the $1\sigma $ standard deviation of the static random error of the single gyroscope, and the compensation accuracy is improved with the increase of the real angular rate modeling order.