Bio-inspired attitude determination method based on polarized skylight and ground horizon

Due to the existence of the gyroscope bias, the attitude obtained by the MEMS inertial sensor easily diverges over time. However, animals in nature can use the natural attributes for navigation, such as geomagnetic field, gravitational field, polarized skylight and stars. The purpose of this paper is to investigate an attitude determination method which mimics animal polarization and omnidirectional vision. This method uses a bio-inspired polarization sensor and a central catadioptric camera to estimate attitude by measuring polarized skylight and ground horizon. We first described the overall framework of the novel bio-inspired attitude determination method and analyzed the measurement principle of polarized skylight and ground horizon. Then, the estimated attitude and MEMS inertial results are fused by a closed-loop Kalman filter. The experimental results indicate that our proposed method can effectively correct the attitude drift of the MEMS inertial sensor. This work provides a novel attitude reference method for potential robot navigation.