2D LiDAR Point Cloud Correction Method for Spherical Robot

To address the distortion of point cloud data caused by the pitch angle oscillation of the LiDAR during the motion of a spherical robot, this study proposes a correction method for LiDAR point cloud data based on IMU. The proposed method achieves time synchronization between IMU and LiDAR data through spherical linear interpolation. A composite criterion combines point cloud distance differences with the LiDAR pitch angle to identify outlier point cloud data. Based on this, a linear regression model is employed to predict and correct these outliers, and coordinate transformations are applied to adjust the pitch angle of the point cloud data, resulting in corrected LiDAR point cloud data that more closely approximates the true values. A spherical robot simulation model was constructed based on Gazebo, and the simulation results demonstrate that the proposed method effectively reduces the error in LiDAR point cloud data and enhances the environmental detection accuracy of the spherical robot.