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

Abstract

This work proposes a fast algorithm for scheduling and trajectory planning of multiple Cartesian robotic arms harvesting fruits in an orchard row segment. The algorithm partitions the workspace adaptively along the segment, computes tight and feasible fruit-picking schedules and vehicle travel speed, and generates smooth, collision-free arm trajectories. The fruit-picking throughput FPT achieved by the algorithm was assessed for a harvester design with up to 15 arms, using two datasets featuring synthetically generated and digitized fruit coordinates, respectively. The algorithm was compared against approaches that partition the workspace into rows of equal height or equal fruit load, and against a relaxed/simpler Mixed Integer Programming (MIP) formulation that ignores collision constraints. The throughput of the proposed algorithm increased monotonically as more arms were added, and its performance was close to that of the MIP, and up to 15.3% and 64.3% better than the equal-load and equal-height workspace partitioning, respectively. Adding more arms when fruit densities were low resulted in diminishing gains, as it took longer to travel from one fruit to another. However, when there were enough fruits, the proposed algorithm achieved a linear speed-up of the FPT as the number of arms increased.