Research On Autonomous Picking Robot Competition Platform System

This paper through research of independent picking robot system of China Robo Cup contest platform to guide the picking robot project improvement and promote the development of contest events to high technology level, to provide better communication platform for college students and more robot enthusiasts, also has important practical significance of promoting robotics applications in the field of agriculture.This paper mainly studied picking robot visual positioning and robotic visual servo positioning control, built a self-picking robot, by combining with requirements of real-time and accuracy of picking robot to identify and locate target fruit in contest, mainly studied three-dimensional position fruit quickly location and manipulator move controling. This papers’ specific research are as follows:First, through the analysis of picking requirements, design the overall structure of picking robots, build a robot hardware platform, based on the structure of the manipulator establish kinematic model to do positive and inverse kinematics analysis. Determine the control scheme of robot system and completed the control system hardware circuit design and produce.Secondly, study the three-dimensional space fruit identification & positioning in the RGB color space find optimal color model threshold segmentation method to achieve the goal of effective segmentation target fruit. Adopts Run-length coding method to do connect region analysis and select the best picking objects. Aimed at target fruit appearance, small texture difference, easy to appear multiple candidate matching cases adopts coarse to fine method. To design fruit picking robot binocular vision positioning system, first adopts centroid feature matching lock candidate matching points. Then use neighborhood gray cross-correlation measures remove the mismatching to improve matching accuracy.Finally study manipulator motion control under the guide of visual system. First according to the binocular vision system measure the coordinates of the target fruit, obtained manipulator target position corresponding joint angles value by using inverse kinematics of manipulator motion model. And in the joint space do track interpolation of manipulator motion trajectory. Then design the part of picking manipulator controller. This paper based on the fuzzy PID controler to realize the robot joint position control,then simulation by simulink.And the research carried out experiments of robot joint debugging and picking, Verified the performance and precision of picking robot system.