Fast Heuristic Scheduling and Trajectory Planning for Robotic Fruit Harvesters with Multiple Cartesian Arms

TL;DR

This work addresses the coupled task and motion planning problem for a multi-arm Cartesian fruit harvester (MARHP) to maximize fruit-picking throughput under deadlock and collision constraints. It introduces a fast heuristic that decouples high-level fruit-picking sequence generation from low-level trajectory and vehicle speed optimization, using vertical-zone clustering and a binary search over vehicle speed $v$ to minimize makespan. Collision-free trajectories are produced by solving a MILP with decision variables for positions, velocities, and accelerations, guided by arm sequences ${oldsymbol{\a}^{rc}}$ and schedules ${oldsymbol{S}^{rc}}$. Experiments on synthetic layouts show monotonic throughput gains with more arms, with linear speedup when fruit density is high; for example, at density $100$ fruits/m$^2$, throughput scales from $0.17$ to $2.21$ fruits/s as arms increase from 1 to 12, harvesting $10{,}000$ fruits in the test.

Abstract

This work proposes a fast heuristic algorithm for the coupled scheduling and trajectory planning of multiple Cartesian robotic arms harvesting fruits. Our method partitions the workspace, assigns fruit-picking sequences to arms, determines tight and feasible fruit-picking schedules and vehicle travel speed, and generates smooth, collision-free arm trajectories. The fruit-picking throughput achieved by the algorithm was assessed using synthetically generated fruit coordinates and a harvester design featuring up to 12 arms. The throughput increased monotonically as more arms were added. Adding more arms when fruit densities were low resulted in diminishing gains because it took longer to travel from one fruit to another. However, when there were enough fruits, the proposed algorithm achieved a linear speedup as the number of arms increased.

Figures (6)

• Figure 1: CAD model of a robotic harvester with multiple Cartesian robotic arms on an orchard platform vehicle; groups of arms operate inside rectangular frames.
• Figure 2: Deadlock scenario: executing a fruit-picking assignment would result in the two arms crossing each other vertically.
• Figure 3: Solution Framework.
• Figure 4: Example of fruit clustering and sequence generation for a robot with 2 cells and 2 arms per cell.
• Figure 5: Example of an orchard row segment with a randomly generated uniform fruit distribution (color varies with depth).