The problem of cooperative localization for a small group of Unmanned Aerial Vehicles (UAVs) in a GNSS denied environment is addressed in this paper. The presented approach contains two sequential steps: first, an algorithm called cooperative ranging localization, formulated as an Extended Kalman Filter (EKF), estimates each UAV's relative pose inside the group using inter-vehicle ranging measurements; second, an algorithm named cooperative magnetic localization, formulated as a particle filter, estimates each UAV's global pose through matching the group's magnetic anomaly measurements to a given magnetic anomaly map. In this study, each UAV is assumed to only perform a ranging measurement and data exchange with one other UAV at any point in time. A simulator is developed to evaluate the algorithms with magnetic anomaly maps acquired from airborne geophysical survey. The simulation results show that the average estimated position error of a group of 16 UAVs is approximately 20 meters after flying about 180 kilometers in 1 hour. Sensitivity analysis shows that the algorithms can tolerate large variations of velocity, yaw rate, and magnetic anomaly measurement noises. Additionally, the UAV group shows improved position estimation robustness with both high and low resolution maps as more UAVs are added into the group.

中文翻译：

---

使用具有测距和磁异常测量的成对通信的协作导航

本文解决了在 GNSS 拒绝环境中一小群无人驾驶飞行器 (UAV) 的合作定位问题。所提出的方法包含两个连续步骤：首先，一种称为协作测距定位的算法，制定为扩展卡尔曼滤波器（EKF），使用车辆间测距测量估计组内每个无人机的相对姿态；其次，一种称为协作磁定位的算法，被表述为粒子滤波器，通过将组的磁异常测量与给定的磁异常图匹配来估计每个无人机的全局位姿。在本研究中，假设每架无人机在任何时间点仅与另一架无人机进行测距测量和数据交换。开发了一个模拟器来评估从航空地球物理调查中获得的磁异常图的算法。仿真结果表明，一组16架无人机在1小时内飞行约180公里后，其平均估计位置误差约为20米。灵敏度分析表明，该算法可以容忍速度、偏航率和磁异常测量噪声的较大变化。此外，随着更多无人机被添加到组中，无人机组显示出改进的高分辨率和低分辨率地图的位置估计鲁棒性。灵敏度分析表明，该算法可以容忍速度、偏航率和磁异常测量噪声的较大变化。此外，随着更多无人机被添加到组中，无人机组显示出改进的高分辨率和低分辨率地图的位置估计鲁棒性。灵敏度分析表明，该算法可以容忍速度、偏航率和磁异常测量噪声的较大变化。此外，随着更多无人机被添加到组中，无人机组显示出改进的高分辨率和低分辨率地图的位置估计鲁棒性。