基于复合地标导航的动平台四旋翼飞行器自主优化降落技术

To address visual loss, big dead zone, and low landing reliability in marker-based localization, a moving platform landing system based on hybrid landmark is proposed. The circular ring tag and two-dimensional tag are combined together to provide precise localization of camera in large space. An estimator based on extended Kalman filter is proposed to estimate the pose of moving platform online. Meanwhile, the unknown measurement bias of encoder is considered in the system model to improve the robustness of estimator under wheel-clip and reduce the calibration error from encoders. Finally, based on the statement estimation, an optimal landing strategy is designed using the minimum Jerk trajectory, and the efficient and stable landing of the quadrotor on the moving platform is realized. To verify the effectiveness of the proposed landing system, several simulation and actual landing experiments are given. The experimental results show that the hybrid landmark has small dead zone and realizes the complete and comprehensive localization from 0.5 m to 6.0 m, the estimator can obtain accurate pose of the moving platform under the unknown measurement bias of encoder, and the proposed optimal landing strategy can realize reliable landing on the moving platform.