Abstract
In this paper, a Model Predictive Control (MPC)-based approach for vineyard spraying is presented, able to adapt to different vine row structures and suitable for real-time applications. In the presented approach, the mobile base moves along a row of vines while the robotic arm controls the position and orientation of the spray nozzle. A reference lawnmower pattern trajectory is generated from the vine canopy description, with the aim of minimizing waste while ensuring vine coverage. MPC is used to compute the trajectory of the vehicle along the row and the manipulator tool trajectory, which follow the spray reference, while minimizing vehicle acceleration and tool displacement. The manipulator tool velocity commands provided by the MPC algorithm are tracked using task space control. The presented approach is evaluated in two experiments: a vineyard spraying scenario and an external evaluation scenario in an indoor environment equipped with the Optitrack camera system.
Highlights
Agricultural robotics is an exciting, emerging research field that offers a potential solution to the problem of increasing global demand for food production due to exponential population growth and labor shortages [1]
Task space control is used to calculate the joint velocity commands for the robot arm, which are not considered in the planning phase (MPC phase) of the algorithm
The mobile base is controlled based on odometry feedback, which may lead to certain reference tracking problems since there is no external sensing
Summary
Agricultural robotics is an exciting, emerging research field that offers a potential solution to the problem of increasing global demand for food production due to exponential population growth and labor shortages [1]. This paper attempts to present a solution to the following problem: given a description of a row of grapevines, one must select coordinated mobile vehicle and robot arm commands that result in satisfactory canopy coverage, while aiming to minimize spraying agent waste, and perform the task as quickly as possible. This kind of problem setup calls for a control method that is able to adapt to different row structures, accelerate in areas of the row without grapevines and slow down in areas with the largest foliage heights
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call