{
  "paper_id": "1ce90bdafb03435bba7051fb6cc8e777",
  "paper_title": "GIF-LIO: Geometry-Intensity Fused LiDAR–Inertial Odometry With Degeneracy-Constrained State Estimation",
  "paper_authors": [
    "Hekai Yao",
    "Xuetao Zhang",
    "Gang Sun",
    "Yisha Liu",
    "Yan Zhuang"
  ],
  "publication_date": "",
  "publication_venue": "IEEE Transactions on Instrumentation and Measurement",
  "abstract": "Robust and accurate LiDAR odometry is crucial for enabling reliable navigation and mapping for autonomous robots in perceptually challenging environments, such as long corridors, tunnels, and open squares. This article proposes a novel geometry-intensity fused LiDAR-inertial odometry (LIO) framework that enables accurate localization in geometrically degenerate environments. Different from the existing methods, the key insight of the proposed method is that the redundancy-degeneracy-aware geometric measurements are fused with complementary intensity measurements to construct degeneracy-optimization constraints. This ensures that the system state is updated along the higher information directions, thus enhancing localization accuracy and system robustness. In particular, the novel geometric measurement model is proposed to minimize redundancy in scan-to-map matches and indicate nondegenerate directions for state optimization. In addition, the probabilistic adaptive voxel map is leveraged for more accurate geometric measurements and degeneracy detection. When the above geometric measurements and the complementary intensity measurements are both available, the joint update of the constrained iterative extended Kalman filter (CIEKF) is performed to obtain the state estimation. Comparative experiments on public datasets are conducted to demonstrate the superior performance of the proposed odometry in terms of localization accuracy and map consistency, particularly in degenerate environments, such as long tunnels and flat fields.",
  "keywords": [
    "Constrained state estimation",
    "geometry-intensity fusion",
    "LiDAR-inertial odometry (LIO)"
  ],
  "paper_institutions": [
    "Dalian University of Technology",
    "Dalian Maritime University"
  ],
  "output_language": "zh",
  "provider": "doubao",
  "model": "glm-5.3",
  "prompt_template": "simple",
  "extracted_at": "2026-09-17T09:40:16.108973Z",
  "template_tag": "simple",
  "summary": "### 摘要\n\n针对自主机器人在长走廊、隧道和开阔广场等感知挑战环境中，LiDAR-惯性里程计（LIO）因几何量测退化导致累积漂移甚至定位失败的问题，本文提出几何-强度融合的LIO框架GIF-LIO。现有退化鲁棒方法分为几何退化感知优化与LiDAR强度信息增强两类，但前者的几何量测存在冗余与不精确、且约束难以无缝嵌入迭代扩展卡尔曼滤波（IEKF）框架，后者受不精确几何量测制约而降低退化检测质量与强度信息效能。本文核心机制是为扫描-地图配准残差设计Tikhonov正则化项 E(x)=Σ‖n_i^T(^G T_L ^L p_i − ^G q_i)‖² + λ_D‖V_t x‖²，其Hessian矩阵 H_L(x)=2λ_D V_t^T V_t 作为信息精炼矩阵注入卡尔曼增益 K=(J^T R^{-1} J + H_L(x)* + P^{-1})^{-1} J^T R^{-1}，确保系统状态沿高信息方向更新。系统由五个模块组成：预处理（IMU前向预测与反向传播运动畸变校正）、信息精炼（基于梯度流熵最大化的冗余最小化点云采样RMS，检测退化方向V_t）、信息增强（沿退化方向选取强度图像特征构建互补光度量测 z^pho=I_c(Π(^L p_f))−i_f）、概率自适应体素地图（建模点量测噪声与位姿误差以提升点-面匹配精度）以及约束迭代扩展卡尔曼滤波（CIEKF）联合更新。实验选用FusionPortable V2（通用环境）、Enwide（典型几何退化环境，含长隧道、跑道、开阔场地）和NTU-viral（用于信息增强模块消融）公开数据集，均由128线Ouster LiDAR采集并提供可靠强度信息，在Intel i9-13900K CPU（5.80 GHz、32 GB内存）的ROS环境下运行，以ATE RMSE和平均地图熵（MME）为指标；结果表明该方法在通用与退化环境中的定位精度和地图一致性均优于Fast-LIO2、Point-LIO、VoxelMap和Coin-LIO（如Enwide的TunnelD为0.339 m、RunwayS为0.540 m，多数序列最优），验证了几何-强度融合与退化约束机制的有效性，且平均每帧0.072 s（约13.9 Hz）满足实时性。未来方向包括引入深度学习提取高判别性点云特征以应对弱几何约束场景，以及发展自适应、场景感知的λ_D调参策略以降低参数环境特异性。该方法适用于IMU提供高频旋转量测、退化主要源于平移的实际场景，部署时需在实时性约束下权衡冗余最小化与体素地图更新的计算开销。\n\n### 关键要点\n\n- 提出GIF-LIO几何-强度融合LIO框架，将冗余-退化感知的几何量测与互补强度量测融合构建退化优化约束，使状态沿高信息方向更新，实现几何退化环境中的精确鲁棒定位。\n- 信息精炼模块采用冗余最小化点云采样（RMS）框架，通过梯度流熵最大化选取信息量最大的LiDAR点用于扫描-地图配准，减少冗余量测并指示非退化优化方向。\n- 为扫描-地图配准残差设计Tikhonov正则化项，其Hessian矩阵 H_L(x)=2λ_D V_t^T V_t 作为信息精炼矩阵调整卡尔曼增益，抑制系统沿无信息退化方向的状态更新。\n- 信息增强模块将点云球面投影为强度图像并消除线状伪影，沿检测到的退化方向V_t选取贡献最大的图像块构建互补光度量测 z^pho=I_c(Π(^L p_f))−i_f。\n- 采用概率自适应体素地图建模LiDAR点量测噪声与位姿估计误差，实现更精确的点-面匹配、几何量测与退化检测，且用原始未畸变点云建图、剪枝点用于配准与退化检测。\n- 实验基于FusionPortable V2、Enwide和NTU-viral公开数据集（128线Ouster LiDAR采集），在Intel i9-13900K CPU的ROS环境下与Fast-LIO2、Point-LIO、VoxelMap和Coin-LIO进行定量与定性对比。\n- 在FusionPortable V2通用数据集上，本方法在多数序列取得最优ATE RMSE（如h-grass00为0.148 m、l-grass00为0.085 m），所建地图保留树木与墙壁细节、模糊与重影伪影更少且MME局部一致性更高。\n- 在几何退化的Enwide数据集上，Fast-LIO2、Point-LIO和VoxelMap在隧道、跑道等严重退化场景失败，本方法在所有序列成功运行并取得最优精度（如TunnelD为0.339 m、RunwayS为0.540 m、KatzenseeD为0.148 m）。\n- 消融实验表明移除信息精炼模块（w/o IR）或概率体素地图（w/o VM）均降低定位精度，验证两模块在退化环境中提升定位精度与地图局部一致性的有效性。\n- 退化正则化参数λ_D敏感性分析（0至100、步长10）显示最优值随环境结构变化而非普适，实验中统一设为20，多数场景在10–30范围内表现良好，未来需自适应调参策略。\n- 系统平均每帧处理时间0.072 s（约13.9 Hz），超过Ouster OS0-128传感器10 Hz扫描率，尽管冗余最小化与体素地图更新引入额外开销（对比Coin-LIO的0.021 s），仍满足实时性要求。\n- 该方法的退化感知建模聚焦于平移退化，主要适用于IMU提供高频旋转量测且退化主要源于平移的场景；LiDAR-IMU外参假设已知且不在线估计，仅检测平移方向的退化。"
}