A Complementary Filter for High-fidelity Attitude Estimation based on Decoupled Linear/Nonlinear Properties of Inertial Sensors

Abstract

A high-fidelity attitude estimation technique is proposed for mobile robots which move irregularly in wide space, where heterogeneous inertial sensors are combined in a complementary way in the frequency domain. While the working frequency ranges of each sensor are not necessarily complementary, inverse models of them compensate the sensor dynamics and enlarge their effective working ranges. The problems to be addressed are that the sensor dynamics displays a highly nonlinear property in the case of 3D rotation, and, even if it is approximated by a linear system, the inverse models of them tend to be non-proper and unstable. An idea is to decouple it into the dynamics compensation part approximated by a linear transfer function and the strictly nonlinear coordinate transformation part. By inserting the designed filter before the coordinate transformation, the total transfer function is guaranteed to become proper and stable. Particularly, the differential operator of a high-pass filter cancels the integral operator included in the dynamics compensation of the rate gyroscope, which causes instability.