FWFR-SLAM: A Robust Monocular SLAM System for Flapping-Wing Flying Robots

View at publisher Summarise with VIRAT
Authors: Zheng Zhong, Zeyang Yu, Lin Dong, Qiang Fu, Xiaoyang Wu, Wei He
Cited by: 0
References: 45

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Simultaneous localization and mapping (SLAM) technology enables real-time and accurate positioning of flapping-wing flying robots (FWFRs) through sensor data such as vision, which is one of the key technologies for autonomous navigation of FWFRs. However, its robustness and accuracy in complex environments are still insufficient. In this paper, we propose a robust monocular visual SLAM system called FWFR-SLAM for FWFRs in complex environments. First, in the front-end feature matching, the two-stage feature matching method of global search and local matching in S-HFNET algorithm is used to improve the matching accuracy. Then, the uncertainty modeling of initialization map points is introduced in the monocular initialization process, and the reliable initialization map points are selected according to the evaluation results to reduce the front-end mismatches and redundant calculation. Furthermore, a keyframe selection strategy combining image information entropy and motion instability energy is proposed, which only retains representative frames. Finally, we conduct experimental verification through open source datasets and self-collected datasets. Quantitative results show that compared with ORB-SLAM3-S (ORB-SLAM3 after image stabilization), FWFR-SLAM improves positioning accuracy by more than 41% in challenging environments.

Fields of Study & Subjects

Robotics and Sensor-Based Localization Aerospace Engineering and Control Systems 3D Surveying and Cultural Heritage

Publication details

Journal / source
IEEE Transactions on Instrumentation and Measurement
Volume
75
Pages
1–14
Publication date
1 January 2026
Language
EN