On forecast modeling of MEMS gyroscope random drift error

MEMS gyroscope has many advantages, such as low cost, small size, low power consumption. However, due to low precision and random drift error, its further application and development is limited. In order to improve accuracy of the MEMS gyroscope, the gyroscope drift data were processed as follows. First, wavelet transform is used to suppress all kinds of interference noise. Second, forecast model is established using three kinds of SVM(support vector machines) methods to predict the drift data. Final, by comparing of the predicted data and the actual data, the accuracy of the model is analyzed. The experimental results show that LS-SVM(least squares-support vector machine) method combined with wavelet can meet the requirement from both model accuracy and speed of modeling.