AC-LIO: Toward Asymptotic Compensation for Distortion in LiDAR-Inertial Odometry via Selective Intraframe Smoothing

Existing LiDAR-Inertial Odometry (LIO) methods typically utilize the prior trajectory derived from inertial measurement unit (IMU) integration to compensate for the motion distortion within light detection and ranging (LiDAR) frames. However, discrepancies between the prior and the true trajectory can lead to residual motion distortions that compromise the consistency of LiDAR frames with their corresponding geometric environment. This imbalance may result in pointcloud registration becoming trapped in local optima, thereby exacerbating drift during long-term and large-scale localization. To this end, this work proposes a novel LIO framework with selective intraframe smoothing dubbed AC-LIO. The core idea is to asymptotically backpropagate the current update term and compensate for residual motion distortion under the guidance of a convergence criterion, aiming to improve the accuracy of discrete-state LIO systems with minimal computational increase. Extensive experiments demonstrate that the AC-LIO framework further enhances odometry accuracy compared to state-of-the-art methods, with an approximately 28.5% reduction in average root mean square error (RMSE) over the second-best result, leading to marked improvements in the accuracy of long-term and large-scale localization and mapping.