Control Techniques in Robotic Harvesting

Food security and global market competition along with reduced manpower due to continuous decline in seasonal labor availability, high insurance costs, and rising wages are the major thrusts that are driving the global research efforts to develop robotic solutions to fruit and vegetable harvesting operations. As we increasingly seek autonomous and robotic solutions, control systems have become ubiquitous in modern agriculture, and the demand for advanced control techniques that offer performance and robustness guarantees is growing rapidly to realize highly efficient autonomous operations. Visual servo control is a popular control technique in robotic harvesting that refers to controlling the motion of a robot using image feedback obtained from one or more cameras. In this chapter, we present the basics of vision-based control and review the progress made in visual servoing in robotic harvesting. Subsequently, we will study the process of controller design by presenting a simple visual servo control system. Stability and performance of control systems, which are the key influencers in determining harvest efficiency, can be affected by uncertainties in the system and environment. In this chapter, we will particularly focus on the issues that a fruit may not be stationary as exogenous disturbances such as wind gusts, fruit detachment forces, canopy unloading, and robot-tree contact may cause unknown time-varying fruit motion. With this motivation, we will present robust and adaptive visual servo controllers that are resilient to uncertain fruit motion. Finally, we state open problems that are aimed at improving harvest efficiency and harvesting speed that could be addressed by improvements in the performance of visual servo control systems.