Lidar motion distortion compensation algorithm based on NDT registration and wheel odometry
Lidar is a ranging sensor widely used in simultaneous localization and mapping(SLAM),and generally col-lects geometric information of the surrounding environment based on the rotation mechanism.When the lidar moves during the scan,the resulting point cloud will have motion distortions,which in turn reduces the accuracy of the SLAM system.In the SLAM algorithm of lidar,in order to make the estimation result of lidar motion more accu-rate,this paper proposes a motion distortion compensation algorithm of lidar based on normal distributions transform(NDT)and wheel odometry.Firstly,the lidar motion is estimated by high-frequency measurements using wheel odometry,which compensates for some motion distortions.Then,an error processing method is designed based on the NDT registration algorithm.By reducing the influence of odometry drift through accurate matching of point clouds,accurate motion estimation of lidar and accurate compensation for motion distortion can be achieved.The proposed algorithm is tested by datasets and real scene experiments.The experimental results show that,compared with the traditional odometry-assisted method,the proposed algorithm can optimize the motion distortion compensa-tion effect,reduce the cumulative error of the trajectory and generate a globally consistent map.
simultaneous localization and mapping(SLAM)motion distortion compensationlidarwheel odometrynormal distributions transform(NDT)