一种无人直升机自主着舰相对位姿估计方法

A relative pose estimation method for autonomous landing of unmanned helicopter

在无人直升机自主着舰中,鉴于视觉方法在机舰相对位姿估计中存在延迟的缺点,采用视觉惯导融合的方法对机舰相对位姿进行了估计。首先,建立了机舰相对位姿估计框架和相对惯导方程,在其基础上构建了基于IMU和视觉信息的状态方程和量测方程;然后,综合考虑视觉和IMU传感器的特点,采用MAEKF算法对视觉和IMU数据进行融合;最后,基于时间戳对量测数据的融合机制进行优化,并简化状态转移矩阵和系统噪声协方差矩阵的更新方式。仿真实验结果表明,该算法具有较高的精度和鲁棒性。

In the autonomous landing of unmanned helicopter,in view of the disadvantages of the vi-sual method in the pose estimation of unmanned helicopter relative to ship,the visual inertial navigation fusion method was adopted for the pose estimation.Firstly,the relative pose estimation frame and relative inertial navigation equation for unmanned helicopter and ship were established,on basis of which the state equation and measurement equation based on IMU and visual information were constructed.Secondly,considering the characteristics of vision and IMU sensors,the MAEKF algorithm was proposed to fuse the visual and IMU data.Finally,the fusion mechanism was optimized based on timestamp,and the updating mode of state transition matrix and system noise covariance matrix was simplified.Simulation results show that the algorithm has high precision and robustness.