| In recent years,with the adjustment of the national economic structure,the trend of population aging has become more and more obvious,and the rural population has been migrating to cities and towns,resulting in a continuous decline in the rural labor force and increasing labor costs.The fruit industry must reduce labor demand through mechanized operations.Apple harvesting is a seasonal,labor-demanding,inter-garden labor.However,whether it is China or other developed countries,most of the methods of moving fruit boxes on standard orchards are still to use manpower or auxiliary platforms to carry them,which is not only inefficient,but also requires high labor.Therefore,the development of fruit box handling equipment with autonomous navigation function can not only improve efficiency,lay the foundation for the construction of unmanned orchards,but also an effective means to solve the shortage of agricultural labor,which is extremely important for the realization of unmanned automatic transportation of fruit boxes in orchards.realistic meaning.The main work of the research is as follows:(1)The overall scheme of the orchard fruit box automatic handling vehicle is designed,the software system mainly introduces the navigation software framework,through LIDAR scanning,obtaining LIDAR data to establish a global cost map,and setting the starting point and end point on the global cost map,the ROS system will automatically generate an optimal path,and the handling vehicle will travel along the path;hardware The system mainly includes the installation of the rear frame of the tractor,the installation of the absolute value encoder of the front wheel steering,the installation of the servo motor of the steering wheel control part,the modification of the rear fork handling device,the modification of the adjustment manipulation mechanism,the installation of the ultrasonic distance measuring module,the installation of the speed measuring device of the handling vehicle and the installation of the Li DAR,IMU and camera,etc.Some corresponding experiments were conducted to test the feasibility of the hardware installation.(2)Complete the study of map construction for orchard fruit box handling equipment.Firstly,the map types were introduced;secondly,the principles of two map construction algorithms were introduced: the RBPF-based Gmapping algorithm and the graph optimization-based Hector＿SLAM algorithm,and a comparison between the two algorithms was conducted.SLAM is more consistent than Gmapping,and Gmapping is more computationally intensive and requires a large number of particles as well as high requirements for computer performance,so Hector＿SLAM is more advantageous in comparison,and Hector＿SLAM is chosen to build the map.(3)Complete the research on information fusion and fruit box detection methods.Firstly,the YOLOV4 model was used for fruit box identification and detection in orchards,and data fusion between Li DAR and camera was carried out to obtain a target identification model applicable to orchard environment by filtering and optimizing VOC data.Secondly,the joint calibration method is introduced,and a black-and-white checkerboard calibration board of A0 size is made using Zhengyou Zhang’s checkerboard grid calibration method,and the corner point information on the calibration board at important locations and the point cloud data on the calibration board acquired by the manually selected LIDAR are obtained several times,and then the software is used to calculate the coordinate relationship between the two,and the corresponding calibration parameters are obtained to realize the camera and LIDAR data fusion.Finally,the joint calibration was tested in a real environment,and the results showed that the error did not exceed 2.5%.(4)Complete the research on path planning and tracking of automatic fruit box handling vehicle in orchard.Firstly,an adaptive Monte Carlo algorithm based on Monte Carlo was proposed,secondly,Dijkstra and A* were compared in the selection of global path planning,and the A* algorithm was selected by simulation comparison on Matlab,and the DWA algorithm was used for local path planning,they were simulated and analyzed on Gazebo simulation software,and the results showed that the method has good navigation performance.For the automatic driving requirement of orchard operation vehicles,a pure tracking and tracing model is used,and the simulation is carried out on ROS system to compare the driving effect of the handling vehicle and determine the best forward-looking distance at different forward-looking distances.The average deviation of the path is the smallest and the tracking quality is the best when the front view distance LD=3m compared with other front view distances;finally,the path planning test is carried out in the real environment and the environment map is obtained,and the orchard fruit box carrier can complete the navigation task better based on this navigation algorithm and ensure that it can avoid obstacles and reach the destination when traveling. |
PDF Full Text Request