基于雷达–扫描器/惯性导航系统的微小型无人机室内组合导航

Indoor integrated navigation of micro aerial vehicle based on radar-scanner and inertial navigation system

本文提出一种基于雷达–扫描器/惯性导航系统(radar-scanner/INS)的微小型无人机室内导航方法.为提高算法的实时性,采用基于扩展卡尔曼滤波(EKF)的D&C同步定位与构图技术(SLAM)实现定位和构图;在更新状态值的扫描匹配过程中提出启发性逻辑来筛选激光雷达数据,以提高算法对无人机因姿态和高度变化而引起的轮廓地图波动的抗干扰性;在特征匹配的过程中选取合理的地图轮廓特征,并利用扫描匹配的结果和特征匹配的传递性提出了精度较高的引导配对,以提高特征配对在三维环境下的准确性;最后,将D&C SLAM与惯性导航系统进行基于EKF的组合滤波,给出无人机的全状态估计.通过与GPS/INS组合导航对比以及室内飞行验证,本文提出的方法能够满足无人机飞行控制对导航实时性和精度的要求.

This paper presents an indoor integrated navigation method for micro aerial vehicles based on radar-scanner and inertial navigation system （INS）. We employ the D＆C simultaneous localization and mapping （SLAM） technique, which is actually an EKF-based SLAM algorithm, to meet the real-time requirements of the micro aerial vehicle （MAV） navigation. For the scan matching in state-update procedure, a heuristic logic is proposed to screen the radar data to deal with the strong perturbations on the map profile caused by the variations in attitude and height during the motion of MAV. We extract reasonable features from the map profile and use them in feature matching procedures. To improve the matching accuracy, we determine the guiding matching in pair-wise based on the result of scan matching and the transfer of feature matching. Finally, an integrated D＆C SLAM and INS navigation algorithm is implemented based on the extended Kalman filter. By comparing the indoor integrated navigation system with the typical global positioning system （GPS）/INS integrated navigation system, we find that the former is capable of providing desired real-time state estimations for MAV＇s flight control, which is further verified by an equipped quadrotor＇s indoor autonomous hover test.