In context
In mid-2025, autonomous drones were increasingly deployed for inspection and emergency response, yet safe landing in unknown terrain remained a critical bottleneck. This work from Xiamen University addressed the challenge of reliable autonomous landing when GPS or visual cues are unreliable, offering an engineering solution that could extend drone operational safety in industrial settings.
What was reported
Researchers developed a LiDAR-based landing system that fuses data from a Livox-Avia solid-state LiDAR and an inertial measurement unit (IMU) using the FAST-LIO2 localization framework. Point clouds are projected into a 2.5D elevation map, which is densified via Bayesian generalized kernel elevation inference and an enhanced dynamic point update algorithm. The update method adds a historical point check to prevent erroneous clearing of objects at the sensor's field-of-view boundary, such as umbrella-shaped vegetation.
Landability is assessed by computing three terrain geometric parameters from the elevation map: slope (via SVD plane fitting), step height (maximum elevation difference among eight neighbors), and roughness (standard deviation of projected elevations). A landability map is generated by thresholding these parameters and weighting them to score each cell. A GPU-accelerated search then identifies the safest landing position in real time.
To handle potential LiDAR localization degradation near the landing point, the system employs a fallback obstacle avoidance strategy using only a single frame of point cloud. Tests in multiple complex simulated environments and real-world scenarios demonstrated effective autonomous safe landing.
Why it mattered
This work advanced the practicality of autonomous drone landing in unstructured environments by improving map accuracy, real-time performance, and robustness to sensor limitations. For industrial automation, it offered a pathway toward safer drone operations in inspection, logistics, and emergency response, where reliable landing is essential for mission success.
“The proposed method achieves excellent autonomous and safe landing results in tests in multiple complex simulation environments and real scenarios.”
Source: 《机器人》期刊 (robot.sia.cn) · Published 2025-06-09 · “基于高程图可着陆性分析的无人机可靠自主着陆与避险方法”