Measurement-Driven Uncertainty-Aware Visual–LiDAR–Inertial SLAM for UAV Navigation in Degraded Environments
Jun Li, Keqin Dou, Jinsong Liu, Hanbin Zhou, Yuejian Chen, Ali Abbas Zaidi, Ke Feng
China Industrial Control Systems Cyber Emergency Response Team Xi'an Jiaotong University University of Manitoba
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
Robust localization in GPS-denied environments requires multisensor fusion that accounts for time-varying measurement reliability. This work presents an uncertainty-aware visual–LiDAR–inertial SLAM framework incorporating learning-based and physics-informed modeling within a unified factor graph. Instead of fixed covariances, sensor contributions are inferred from information-matrix traces, enabling automatic adaptation to environmental and kinematic degradation. Evaluation across three controlled scenarios demonstrates a mean absolute trajectory error (ATE) of 0.167 m, outperforming deterministic fixed-weight fusion. Calibration analysis confirms statistically consistent uncertainty estimates ($\boldsymbol {R}^{\mathbf {2}}\ge \mathbf {0}.\mathbf {986}$, expected calibration error (ECE)$\le \mathbf {0.\mathbf {021}~\mathbf {m}}$), highlighting the framework’s effectiveness.
逐年被引趋势
暂无年度引用数据
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
计算机 / AIRobotics and Sensor-Based Localization
Inertial Sensor and Navigation · Target Tracking and Data Fusion in Sensor Networks