Research On Bionic Design And Operation Performance Of Tomato Picking Robot Claw

At present,artificial harvesting of tomatoes is still the mainstream,with high labor intensity,low efficiency and insufficient farming population.Therefore,there is an urgent need to employ robotic gripper for tomato picking.Most of the research results of domestic and foreign tomato picking robotic grippers are picked by gripping,twisting,adsorption and other methods,which have problems such as low success rate and potential fruit damage.In this study,small-sized tomato varieties are used as the picking object,and an enveloping picking robotic gripper with an integrated visual feedback function is designed by referring to bionic ideas.Through numerous picking experiments,the various operating performances of the designed envelope-type picking robotic gripper are verified.The specific research contents are as follows:First,the physical characteristics of the target tomato,including the structural size,skin friction characteristics,and pressure characteristics,are studied to provide data support for the design of the robotic gripper.Three commonly used tomato machine harvesting methods are compared to identify their advantages and disadvantages.Based on this comparison,the envelope-picking method is selected,which clarifies the direction for the subsequent design of the robotic gripper structure.Secondly,referring to the advantages of the single-toe structure of the cat’s front paw,one arc-shaped single-toe structure and two single-toe driving mechanisms are designed.The structure of the single-toe is optimized through AN SYS statics simulation analysis.The kinematics analysis and ADAMS simulation of the two single-toe driving mechanisms are carried out respectively,and the chute mechanism with better performance is selected as the enveloping claw single-toe driving mechanism.6 single-toes are evenly distributed in the circumferential direction to form an enveloping claw.In order to assist the picking of the enveloping claw,an auxiliary picking tool imitating the mandible is designed.After establishing and analyzing the contact model between the target tomato and the knife,it is concluded that a distance of 0.88mm between the outer surface of the trough and the outer surface of the blade,as well as a distance of 0.88mm between the top of the trough and the top of the blade,will ensure the blade does not touch the target tomato.The auxiliary picking tool and the enveloping claw together form the core structure of the robotic gripper.A three-dimensional model of the overall structure of the robotic gripper is established,and a prototype is fabricated.Then,a closed-loop control strategy for picking operations is proposed and the visual feedback function based on the YOLOv5s algorithm is realized.The precision of the detection model reaches 99.92%,the recall rate reaches 99.89%,and the mean average precision reaches 94.63%.The intelligent feedback function of the robotic gripper has been implemented,preventing it from being unaware that it did not successfully pick up an object.Finally,a test platform is built and utilized to verify the picking performance of the robotic gripper designed in this paper.The experimental results show that:the auxiliary picking tool imitating the mandible can reduce the maximum pulling force required for successful tomato picking by 43.5%;the envelope claw of the robotic claw allows for a positioning error within 12mm when picking tomatoes;the visual feedback function achieves a detection success rate of 98.6%in actual work;using the picking robotic claws to carry out picking experiments on sparsely distributed and densely distributed tomato plants,the direct damage rates of the tomatoes are 2%and 0%respectively,the qualitative change rates of tomatoes 72h after picking is 8%,6%respectively,the picking efficiencies are 15.01s/piece and 14.54s/piece respectively,the visual feedback function increases the picking success rates by 8%and 6%respectively,reaching 90%and 96%.In summary,this thesis has successfully developed an envelope-type picking robotic gripper for tomato picking to address the issues of existing picking robotic grippers and tomato-picking operations.Through experiments,it has been shown that the developed picking robotic gripper causes less damage to the target tomatoes,has high picking efficiency with a success rate exceeding 90%,and performs well in tomato-picking operations.It can effectively complete the task of tomato picking.