Open AccessModeling the Configuration of a Robotic Gripper for Handling Agricultural Products
Author(s) -
Dinh-Tien Vu,
Andrey Ronzhin
Publication year - 2021
Publication title -
izvestiâ ûgo-zapadnogo gosudarstvennogo universiteta
Language(s) - English
Resource type - Journals
eISSN - 2686-6757
pISSN - 2223-1560
DOI - 10.21869/2223-1560-2020-24-4-76-90
Subject(s) - kinematics , software , robot , computer science , control engineering , engineering , simulation , artificial intelligence , classical mechanics , programming language , physics
Purpose of research. Improvement of quality and speed of harvesting agricultural products through the development of models, control algorithms and multi-criteria optimization of the robotic gripper configuration. Methods. To achieve this goal, we have used the methods of mathematical and computer modeling, multi-criteria optimization, the theory of object-oriented design and programming. The mathematical model of the kinematic scheme of the prototype of the robotic gripper, its geometric constraints and objective functions used for optimization are described. Results. It has been performed a review of approaches to robotic harvesting of agricultural products, confirming the relevance of this study of robotic gripper configurations, which provides reliable fixation of an object without causing damage. The results of experiments on evaluating the developed algorithms and a software system for optimizing the configuration of a robotic gripper are presented. The developed software system AgroGripModeling for modeling the configuration of a robotic gripper using three a posteriori algorithms NSGA-II, MOGWO and MOPSO for multicriteria optimization is tested in the design of a prototype of a four-fingered gripper with a vacuum bellows for picking tomatoes. Conclusion. When designing a robotic gripper, it is necessary to take into account the variety of manipulated objects, the complexity of their identification and guidance of the manipulator in a complex natural environment with obstacles. The task of optimizing the capture mechanism is associated with the fulfillment of a number of conflicting requirements for reliability, softness, accuracy, speed, energy efficiency, which form a complex space for finding solutions. The developed AgroGripModeling software system provides modeling of the robotic gripping configuration and its quality assessment using three a posteriori algorithms NSGA-II, MOGWO and MOPSO. The system was tested with multicriteria optimization of the configuration of a prototype of a four-fingered gripper with a vacuum bellows for picking tomatoes.