Scholarlink

Kalman filtering is indispensable in UAV’s vision-based autonomous landing. However, Extended Kalman Filter(EKF) currently employed is increasingly unable to meet the rising requirements. EKF lacks in convergence performance and robustness. To address this issue, we introduces a more effective nonlinear Kalman filter, Invariant extended Kalman Filter (IEKF), into the step of tracking and locating cooperative target. We conducted simulation with Gazebo to verify the improvement of landing performance using IEKF. The use of IEKF has reduced the positional deviation by 85.28% and landing orientation by 41.15% compared to extended Kalman Filter(EKF). To further validate the conclusion, the real flight validation was also conducted.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Vision-Based UAV Precision Landing via Invariant Extended Kalman Filtering

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yiqi Fu,
  • Qingqing Dang,
  • Wen Zhao,
  • Liqiao Li,
  • Zhu Zhu

摘要

Kalman filtering is indispensable in UAV’s vision-based autonomous landing. However, Extended Kalman Filter(EKF) currently employed is increasingly unable to meet the rising requirements. EKF lacks in convergence performance and robustness. To address this issue, we introduces a more effective nonlinear Kalman filter, Invariant extended Kalman Filter (IEKF), into the step of tracking and locating cooperative target. We conducted simulation with Gazebo to verify the improvement of landing performance using IEKF. The use of IEKF has reduced the positional deviation by 85.28% and landing orientation by 41.15% compared to extended Kalman Filter(EKF). To further validate the conclusion, the real flight validation was also conducted.