Research On Autonomous Navigation Of Orchard Vehicles Based On LiDAR

Autonomous navigation and control technology is of great significance to the development of precision agriculture,can effectively solve the current situation of shortage of productivity,and play a significant role in improving the level of mechanization and production efficiency.In recent years,my country’s industrial structure has been continuously optimized,and the country has vigorously implemented innovative development strategies,which have enabled autonomous navigation technology to obtain financial support and become a hot spot in the field of precision agriculture research.With the improvement of people’s living standards,the fruit industry has become an important part of my country’s national economy.However,the degree of mechanization of orchards in my country is low,and the degree of automation of fruit tree planting,management and harvesting is low,which is not conducive to the further development of the orchard industry.In order to improve the automation level of the orchard industry and realize the sustainable development of the orchard,the introduction of autonomous navigation technology into the orchard operation task has become an effective way to solve this problem.In this paper,the standardized orchard is used as the research environment,and the two-dimensional Li DAR is used as the only environmental measurement sensor to obtain the environmental measurement data of the orchard.The main work and conclusions of this paper are as follows:(1)A crawler trolley is built to simulate the actual orchard operation vehicle.The vehicle is divided into upper and lower computers.The upper computer includes a two-dimensional laser radar and a navigation PC.The lower computer consists of a motor controller,a motor drive module,a mobile chassis and it is composed of bluetooth module.Using two-dimensional Li DAR to obtain the measurement point cloud data of fruit trees.The measurement data is processed by the navigation PC to generate the navigation path,calculate the navigation parameters,and transmit the navigation parameters to the navigation controller;the navigation controller receives the navigation control parameters and generates the navigation control decision-making,the motor is driven to rotate by the motor drive module,so that the vehicle can drive autonomously along the navigation path.(2)Aiming at the problems of vehicle yaw and non-adjacent tree-row interference in the process of autonomous navigation between orchards,a Li DAR-based autonomous navigation control system between rows of orchards was designed.Setting a circular area of interest to avoid the interference of non-adjacent tree rows,using detection statistics to remove non-trunk points in the area of interest,and using K-Means algorithm to divide the measured trunk points into two rows,left and right,and then divide the two sides.The tree row is fitted as a tree row straight line,the center line of the tree row is extracted as the navigation path,and the incremental PD controller is used as the navigation controller to make the vehicle travel along the navigation path.The test shows that when the initial lateral deviation of the vehicle is not greater than 50 cm,the deviation between the vehicle trajectory and the navigation path can be controlled within ±3cm within the range of 2m of vehicle travel.(3)On the basis of the autonomous navigation control system between rows of orchards,an algorithm for generating and tracking paths between rows of orchards and a vehicle headland turning strategy are designed.The elliptical region of interest is used instead of the circular region of interest,and the median filtering method is used to weaken the measure of error,a two-step treerow segmentation method is proposed.The pure tracking algorithm is used to track the navigation path.When the vehicle reaches the headland,a turning strategy is generated,so that the vehicle enters the next row and continues to work.In order to test and evaluate the overall autonomous navigation algorithm of the orchard more conveniently,a simulation platform for the overall autonomous navigation of the orchard is built by establishing the orchard planting model,Li DAR measurement model,vehicle kinematics model and motor control model.Navigation path and vehicle motion trajectory,vehicle deviation visualization.The intra-row simulation test shows that when the orchard has a planting deviation of ±20cm and the tree-missing rate is not more than 25%,the orchard has a growth deviation of ±3cm,the initial vehicle heading deviation is less than 15°,and the lateral deviation is less than 1m,the orchard will be driven.The autonomous navigation algorithm can control the lateral deviation within ±17cm.The single-factor simulation test of the overall autonomous navigation of the orchard shows that when the number of missing trees in the orchard is 3,the amplitude of the fruit tree curve is 1,the initial heading deviation of the vehicle is30°,and the initial lateral deviation of the vehicle is 30 cm,the mean lateral deviations are 12.03 cm and 9.07 cm,5.39 cm,and 3.85 cm,respectively.and with the increase in the number of missing trees in the orchard row,the amplitude of the fruit tree curve,the initial vehicle heading deviation,and the heading deviation,the mean lateral deviation during vehicle navigation also increases.