Research On Grasping Of Apple Picking Robot Based On Position And Attitude

Agricultural robot is one of the important equipments of precision agriculture in the 21st century and the development direction of intelligent agriculture machine in the future. Based on the development of the fruit picking robot, we developed the algorithm of apple position estimation, and the algorithm of apple attitude estimation in this paper. The main research contents and conclusions are the following:1. Particle filtering theory was introduced in detail. The inverse kinematics solution of Motoman SSF2000 manipulator were completed and the simulation experiments are executed by OPENGL2. A new position estimation theory was developed. Through fusing the stereo information and monocular information we completed the target position estimate by particle filtering.3. A method of measuring apple attitude was put forward with the machine vision. And two models are divided through the difference of the remains of apple calyx, if the apple calyx can be seen then the apple attitude can be measured by first model, if not, the apple attitude can be measured by second model. Finally, the attitude data was fused with the particle filter, and the optimal estimation of the apple attitude was obtained.4. Picking experiments were executed with apple position and attitude information on the apple picking robot. First we acquired apple's precise position and attitude information by position estimation and attitude estimation, and then completed the grip planning by attitude information and picking requirement.The proposed algorithm not only provided a viable way for apple picking robot design, but also satisfied the requirement of precision pick based on position and attitude.