Random Error Modeling Method for Gyrowheel Based on Empirical Mode Decomposition and Deep Neural Network

Gyrowheel is a promising device for small spacecrafts, as it integrates both the attitude actuating and the angular velocity sensing capabilities. Random error is the major determinant of the measurement accuracy of the gyrowheel. However, large tilt angles and de-tuned spin speeds of the gyrowheel rotor result in special characteristics of its random error, which brings challenges of extracting and modeling the random error. To overcome the challenges, a random error extraction method is proposed by adopting the empirical mode decomposition (EMD) technique. And on this basis, a deep neural network-based random error modeling method for the gyrowheel is developed by utilizing a stack-block architecture. Validation results demonstrate the effectiveness of the proposed method.