融合速度估计与里程计位姿的扫描匹配改进算法

An Improved Scan-matching Algorithm Based on the Fusion of Velocity Estimation and Odometry

为改进低帧率激光雷达在扫描匹配中的扫描畸变问题,提出了融合速度估计与里程计位姿估计的扫描匹配改进算法。首先以迭代最邻近点匹配为基础算法,分析了现有算法的问题所在。其次对扫描畸变问题展开分析,运用里程计位姿估计对激光位姿进行分段位姿估计。通过对分段内激光点进行速度估计,进一步确定激光点位姿。给出了具体的算法原理,并进行了仿真实验和具体的对比实验。仿真结果与实际实验结果表明,本文算法在匹配精度上优于ICP算法及VICP算法。

In order to solve the scan distortion problem in scan-matching algorithm with the low frame rate Lidar,an improved scan-matching algorithm based on the fusion of velocity estimation and odometry was presented.Firstly,based on the iterative closest point(ICP)algorithm,the problem of existing algorithm was analyzed.Then,the scan distortion problem was determined and the point clouds were segmented by odometry.By estimating the velocity of the lidar,the pose of the point was further confirmed.The specific algorithm was given,and the simulation experiment and specific contrast experiment were carried out.Consequently,the simulation experiment and actual experiments show that the proposed algorithm is better than origin ICP and VICP algorithm.