A HARDWARE-ORIENTED METHOD OF ACCELERATED SEARCH BY TEMPLATE BASED ON STRUCTURAL-PROCEDURAL COMPUTING
Evgeny A. Titenko, Эдуард Игоревич Ватутин, М. А. Titenko, А. P. Loktionov, Eduard V. Melnik · Известия Южного федерального университета. Технические науки · 2024
The aim of this work is to develop a trajectory planning system for the movement of a delta robotused for weed cultivation. The delta robot is mounted on a mobile platform that moves between rows ofcultivated plants. A vision system detects weeds and determines their coordinates. The system is taskedwith planning the trajectory of the robot's gripper during weed removal, ensuring no damage is done toeither the robot or the plants. This research is highly relevant due to the growing global population, decreasingarable land, rural depopulation, and a reduction in the availability of agricultural machinery.To achieve this goal, the work presents a solution to both the forward and inverse kinematics of the deltarobot using an analytical approach. A model for determining the structural parameters of the delta robotis proposed, which allows the evaluation of how these parameters affect the robot’s working area.The lengths of the delta robot's arms are determined, tailored to the task of weed removal in corn fields.The trajectory planning problem is addressed by decomposing the motion into horizontal movement of thegripper and vertical movement, considering the size of soil clumps and the magnitude of weed extraction.Experimental results demonstrate the possibility of significantly reducing the number of trajectory points,thus lowering the computational complexity of the proposed methods and simplifying their implementationin the robot's onboard computer.