AS-LIO: Spatial Overlap Guided Adaptive Sliding Window LiDAR-Inertial Odometry for Aggressive FOV Variation

TL;DR

AS-LIO tackles aggressive motion and FOV variation in LiDAR-Inertial Odometry by introducing a real-time Spatial Overlap Degree (SOD) computed via a 3D soft-margin voxel map. SOD directly quantifies the impact of FOV changes on frame-to-map alignment and guides an adaptive sliding window that adjusts the update frequency; historical frame constraints within the window maintain robust data association. The system uses an ESKF-based state estimator with LiDAR-IMU fusion and plane-based LiDAR residuals to produce accurate, high-frequency odometry while suppressing nonlinear errors. Experiments on diverse indoor and outdoor datasets show AS-LIO outperforming state-of-the-art LIO methods in both accuracy and robustness, especially under aggressive turns and challenging occlusions. The approach offers a practical, real-time solution to improve LIO reliability in harsh operational conditions.

Abstract

LiDAR-Inertial Odometry (LIO) demonstrates outstanding accuracy and stability in general low-speed and smooth motion scenarios. However, in high-speed and intense motion scenarios, such as sharp turns, two primary challenges arise: firstly, due to the limitations of IMU frequency, the error in estimating significantly non-linear motion states escalates; secondly, drastic changes in the Field of View (FOV) may diminish the spatial overlap between LiDAR frame and pointcloud map (or between frames), leading to insufficient data association and constraint degradation. To address these issues, we propose a novel Adaptive Sliding window LIO framework (AS-LIO) guided by the Spatial Overlap Degree (SOD). Initially, we assess the SOD between the LiDAR frames and the registered map, directly evaluating the adverse impact of current FOV variation on pointcloud alignment. Subsequently, we design an adaptive sliding window to manage the continuous LiDAR stream and control state updates, dynamically adjusting the update step according to the SOD. This strategy enables our odometry to adaptively adopt higher update frequency to precisely characterize trajectory during aggressive FOV variation, thus effectively reducing the non-linear error in positioning. Meanwhile, the historical constraints within the sliding window reinforce the frame-to-map data association, ensuring the robustness of state estimation. Experiments show that our AS-LIO framework can quickly perceive and respond to challenging FOV change, outperforming other state-of-the-art LIO frameworks in terms of accuracy and robustness.

Figures (6)

• Figure 1: Our AS-LIO framework (Left) and result comparision (Right). We utilize the frame-to-map spatial overlap degree to assess the impact of FOV change on pointcloud alignment, then accordingly adjust the update step of the sliding window and maintain the history constraints.
• Figure 2: The system overview of our AS-LIO framework.
• Figure 3: Our data collection device, including the solid-state LiDAR Livox Avia with the built-in BMI088 IMU, and three other cameras.
• Figure 4: Spatial overlap degree evalutation. We conduct a sensitivity analysis on the SOD computation, indicating SOD is stable for different voxel size (within reasonable ranges). We also compare IMU data within the same period, indicating SOD with a higher signal-to-noise ratio can significantly identify when FOV change weaken pointcloud constraints.
• Figure 5: Odometry accuracy evaluation. The figure shows the result of outdoor_5. We complete a strict closed-loop path (about 400m) in campus environment with enhanced sharp turns at crossings. The odometry trajectory and the consistency of map indicate that our AS-LIO achieves better performance.