Data Fusion Estimation of Inertial Sensors Based on Multiscale Stochastic Dynamic Models

In the paper, combing with discrete wavelet transform, dynamic system theory and stochastic process theory establish multiscale stochastic dynamic models considering scale as variable and present multiscale fusion estimation algorithm in order to realize the optimum estimation of the state. The algorithm may be a method used in no state model. Using the algorithm for gyro signals processing and fusing the observation at different scales, the accuracy is improved obviously. Simulation and test all prove that the algorithm is available.