基于LiDAR多维点云优化的垄作菊花采摘机器人自主导航方法研究OA北大核心CSTPCD

[Objectives]In response to the problems of unstable walking and low production efficiency of the golden chrysanthemum picking robot caused by the complex field operating environment,this paper designed an autonomous navigation system for the chrysanthemum picking robot based on LiDAR multi-dimensional point cloud optimization to achieve precise operation and efficient production of the robot in farmland.[Methods]The tracked chassis was equipped with a Velodyne 16-line LiDAR which was used to obtain 3D point cloud information in the field and coordinate correction and voxel filtering preprocessing were performed on it.A multidimensional point cloud optimization algorithm was proposed,which could adaptively obtain effective point cloud features under different coordinate axes on the characteristics of the golden chrysanthemum plant,and obtain ridge lines on both sides.The final navigation baseline was obtained using the least squares method to fit ridge lines,which was tracked by using an improved pure tracking control algorithm.[Results]Through simulation experiments on the Stanley control algorithm and the improved pure tracking control algorithm,the improved pure tracking algorithm showed more efficient tracking performance.A field experiment was conducted using the golden chrysanthemum harvesting robot at the Hushu Chrysanthemum Garden in Nanjing.The experimental results showed that the autonomous navigation algorithm based on LiDAR multi-dimensional point cloud optimization had a lateral average absolute error of 0.047 4 m,a standard deviation of 0.030 6 m,an extreme absolute position deviation of 0.078 9 m,a heading average absolute error of 2.177°,and a standard deviation of 2.589°,the lateral average absolute error was reduced by 67.13%and the degree of dispersion was reduced by 48.34%.[Conclusions]The LiDAR multi-dimensional point cloud optimization algorithm and improved pure tracking algorithm proposed in this article can effectively improve navigation accuracy,improve system anti-interference,and achieve good navigation results,thereby ensuring the precise operation of the golden chrysanthemum picking robot.