Behavior Planning for Pruning with Robotic Arms

Pruning is an essential agricultural practice for orchards. Properly pruning an orchard can promote healthier growth and optimize fruit production over the lifespan. Robotic arms are developed as an automatic solution for such a repetitive task, which typically requires seasonal labor with specific skills. While previous works mainly addressed the challenges of perception, we focus on the planning problem in pruning. Specifically, pruning poses a motion planning problem within environments that have complex collision issues. In this work, we formulate the planning problem for a high-dimensional robotic arm in the pruning scenario and discuss the potential solutions. In the experiment, we demonstrate that more comprehensive planning methods can enhance the performance of the robotic arms in the pruning task. As a result, this motivates future research and development on planning in pruning.