Research On Headland Turning And Localization For Automatic Robot Operating In Orchard

Our country’s fruit production and planting area are in forefront of the whole world,the fruit economy is rather significant in the modern national economy.Nevertheless,it has only a very short history on orchard machinery automation,and compared with the field environment,the orchard environment is more complex.This situation will lead to a slower development of mechanical automation in the process of production management,and the labor force cost increased every year.To develop the orchard automation and improve the working efficiency and reduce the manpower cost is imminent,thus,there is no doubt that it is an ideal solution for developing robot operating in orchard with self-positioning and automatic navigation.This paper uses the own developed experimental platform of orchard robot and conducts geometric feature map including local and global ones in orchard environment based on GPS and laser radar.In order to realize positioning and automatic navigation in global map and carried out the headland turning and positioning based on the unilateral fruit trees before into the next row.The main achievements and conclusions are shown below:(1)When the robot moves into the dense planting fruit tree rows,a laser radar sensor is used to scan relatively cylindrical trunk.Adopt algorithm based on Euclidean distance to collect data clustering and classification and establish laser radar local coordinate system and extract the center of the trunk under local coordinate system which would be transformed to the world coordinates of GPS.Build out the geometrical characteristics map of the global orchard,which is used for navigation in the next step.Fruit trees under global coordinate system can be obtained intuitively.(2)According to the global geometric feature map which is built,the coordinate values of the trees at the end of the row can be confirmed.In order to solve the localization and turning problems of robot which operating in orchard,the research proposes a method by comparing the robot pose and the coordinate values of the tree at the end of the row.The turning radius can be extracted and the method of guiding robot turn autonomously based on GPS.(3)Before the robot which operating in orchard entering the next new row.self-positioning can be corrected based on unilateral trunks which had been saved previously in the world coordinate system.Using VC++6.0 programming language to complete the PC program.Experiments were carried out in orchard-like environment in campus.Experimental results show that the self-position of robot can be corrected based on the unilateral trunks.