Research And Design Of Vineyard Autonomous Navigation Platform Based On LiDAR And RTK

With the continuous loss of agricultural employees and the increasing labor costs,agricultural production is facing great challenges,especially the orchard planting industry with many operations and labor-intensive.According to incomplete statistics,the labor expenditure in the orchard production process accounts for more than 60% of the total orchard production expenditure,which has become an important factor restricting the development of China orchard industry.The production links such as weeding,fertilization,spraying,pruning,picking and transportation in orchards have the problems of simple equipment,low degree of automation and intelligence,heavy work and high production costs.In order to solve the above problems,it is urgent to study the theory of intelligent operation machinery suitable for the current situation of orchards in China,and to develop corresponding intelligent operation machinery for orchards.However,in the densely planted orchard environment,due to the occlusion of trees,the satellite positioning signal wave is refracted,diffracted and reflected,forming a multi-path effect,which in turn leads to the absolute positioning with the Global Navigation Satellite System(GNSS)as the core.The problem of poor or even unusable methods has led to the slow development of orchard robots.How to extract the robot navigation path and locate the relative position of the robot in the densely planted orchard environment where the GNSS signal is missing has become one of the scientific issues affecting the development of the orchard robot industry.In this paper,aiming at the bottleneck problems such as unstable and incoherent autonomous navigation operation caused by the absolute position error or large deviation caused by GNSS signal reflection and refraction in densely planted orchards,3D Li DAR(3D-Li DAR)and The method of combining RTK combined positioning,combined with SLAM(Simultaneous Localization and Mapping)mapping,orchard tree row assisted positioning and other algorithms,designed and implemented an autonomous navigation platform suitable for standardized planting orchards.The main research contents of this paper mainly include:(1)Design and manufacture a mobile platform for autonomous navigation of vineyards.The platform includes a crawler-type differential chassis and an autonomous navigation system based on an Nvidia artificial intelligence processor(Nvidia Jetson NANO).The system mainly includes an Nvidia artificial intelligence processor,a GNSS-RTK receiver,and a 3D li DAR.Among them,the NVIDIA artificial intelligence processor is used to run the Ubuntu18.04 operating system and the Robot Operating System(ROS)robot operating system.The GNSS-RTK receiver is used to receive multi-mode satellite signals and output the absolute position of the mobile platform’s geodetic coordinate system WGS-84.3D Li DAR,obtains 3D point cloud data of orchard environment,used for SLAM mapping and effective extraction of orchard tree rows to realize auxiliary navigation.(2)Orchard environment map construction and 3D lidar environment perception fusion navigation and positioning algorithm.The algorithm includes the construction of a 3D point cloud map of the orchard;The fusion of the absolute position of the GNSS-RTK receiver based on Kalman Filter(KF)and the relative positioning data of 3D point cloud perception;The perception and recognition of vine rows based on 3D Li DAR point cloud target detection Algorithm,can realize online target detection and clustering segmentation,object extraction and perception expression;Based on RANdom SAmple Consensus(RANSAC)and LSM fruit tree row extraction and detection,headland identification and detection algorithm,can fit tree row straight line,steering curve,and combined with the filtering algorithm to construct the navigation reference line equation.(3)Path planning and trajectory tracking algorithm for autonomous navigation platform.Based on the ROS robot operating system and the developed autonomous navigation system,initialize positioning,task analysis,and generate global and local paths.Use the Stanly control algorithm to track the path,obtain the lateral deviation and vertical deviation of the autonomous navigation operation path,and further optimize the navigation path.(4)Test verification of the orchard autonomous navigation platform.It mainly includes the GNSS-RTK absolute position positioning test verification;The tree line fitting in the vineyard and the headland turning curve fitting test;The object detection test and the perception map generation test;Finally,the tree line straight line fitting and route following integration test are carried out.The test results show that the autonomous navigation platform designed and implemented in this paper is suitable for the orchard environment,and the autonomous navigation operation accuracy within 10 cm of lateral deviation can be achieved.