Design And Research Of Apple Picking Robot With Multiple End-effectors

So far the domestic fruit picking operations are basically manual.In the fruit production process,the picking session is the most time-consuming and labor-intensive,and the cost is high.As the increasingly serious aging problem arises and the agricultural labor is transferred to other industries in our society,the problem of agricultural labor shortage is more and more prominent,and it has become increasingly difficult to put in a lot of labor in the fruit picking process.Therefore,the research and development of fruit picking robot for fruit picking operation are of great significance.Most of the fruit picking robots have only one end-effector at present,the picking of each fruit requires both the robot arm and the end-effector to move at the same time,the picking period of single fruit is longer,and picking efficiency is low,the practical use of picking robot is seriously affected.Aiming at the problem of long picking period and low efficiency of picking robot with single end-effector,on the basis of the national "863" project "the key technology research on mobile fruit picking robot",the design and research of apple picking robot system with multiple end-effectors are carried out in the paper,which can provide reference for improving the working efficiency of picking robot and realizing the practical use of picking robot.The main contents and results of this paper are summarized as follows:(1)Based on the study of the physical properties of apple and citrus,the requirements of structure and function of the end-effector for multiple end-effectors simultaneous and continuous picking operation were put forward,and the end-effectors of cup-shaped and cylinder-shaped structure were designed.The working principle of the two end-effectors was described,the physical prototype of the cylindrical end-effector was developed and the picking experiment was carried out.The cylindrical end-effector was compact and simple to drive,and the non-clamping structure was used to avoid the calculation and setting of the clamping force.The cylindrical end-effector could absorb the error of the end position of the manipulator,and it was suitable for the automatic harvesting of spherical fruit such as apple and citrus,and it could meet the requirements of continuous picking operations.(2)The multi-terminal fruit picking manipulator was designed.The performance characteristics of picking manipulator at home and abroad wre analyzed,and the design requirements of picking manipulator were summarized.Based on the investigation and analysis of the shape characteristics of apple crowns under dwarf and dense planting mode,and the growth and distribution characteristics of fruit on crown,a master-slave structure of multi-terminal picking manipulator was designed.The target horizontal working space of the manipulator was 0-2.5m and the vertical working space was 0.75?2.85m.In the Matlab software environment,the Monte Carlo method was used to simulate the working space of the manipulator,the simulation results showed that the designed multi-terminal manipulator could meet the picking requirement of one side of a fruit tree.(3)According to the characteristics of manipulator with multiple end-effectors,a partitioned fruit tree picking operation strategy was formulated.One side of the whole fruit tree was divided into several picking areas,arid each picking area corresponded to a picking point.In the picking operation,the main arm was moved to the picking point first,and then the end-effectors were assisted by each slave arm to carry out the continuous picking of all fruit in the area,and the fruits were collected collectively.The calculating formulas of the picking point coordinates were deduced by using the principle of "more-moving of master-arm and less-moving of slave arm".On this basis,the kinematics mode for robot manipulator was established,the kinematics equations were derived based on the D-H method.The kinematics simulation of the manipulator was conducted using Matlab Robotics Toolbox.(4)The recognition and localization methods for apple were studied.The requirements of binocular vision image acquisition for partitioned fruit tree picking were analyzed.The calculation method of object distance was deduced,and the object distance was determined by the binocular vision parameter and the arm structure parameter.The method of combining color,area and shape features was used to identify apple.The color difference component R-G of the RGB color space was extracted and the image segmentation was carried out by the Otsu automatic threshold method.The segmented image was denoised by morphological operations,the center coordinates of the fruits were obtained by calculating the fruit area centroid.The image matching method of combining centroid matching with region matching were used to realize stereo match of target fruit,and the binocular parallax method was used to locate the fruit.(5)The recognition and location of oscillating fruit were studied based on monocular vision and ultrasonic testing.Firstly,the acquired sequential images of oscillating fruit were segmented by Otsu algorithm based on R-G color component,morphological operations were employed to eliminate residual noise,the region of target fruit was separated from backgrounds.Secondly,the region of target fruit was filled with gray threshold.Thirdly,the processed sequential images were superimposed and a composite image was obtained,the two-dimensional centroid coordinates of the oscillating fruit were extracted from the fruit oscillating region in the composite image.Fourthly,the picking manipulator was controlled to move and its end was be directed to two-dimensional centroid coordinates,at the same time,the depth information of the target fruit was acquired through ultrasonic testing.The ultrasonic echo signals were extracted,wavelet denoising and Hilbert transform were used to obtain the peak to peak value of the signal.The peak to peak value could be used to realize the recognition of apple fruit and fruit tree branches and leaves.Finally,the end-effector of picking manipulator started to grab the fruit when it was within valid range and the manipulator stopped motion.The experiments under laboratory environment results showed that the success rate of picking could reach to 86%.(6)Based on the main idea of the open robot control system,an open,distributed multi-terminal picking robot control system was designed.The two-layer structure including the upper machine and lower machine was adopted in the robot control system,and the lower machine comprised a master arm controller and slave arm controllers.The slave arm joints were driven by stepping motor,and multi-axis subdivision-driven stepper motors constituted a motion control system.The physical prototype of the multi-terminal apple picking robot system was manufactured,and the picking experiments were carried out under laboratory environment.The results showed that the picking success rate could reach to 82.14%,and the average time of a fruit picking was 4.5s.