Investigation on the Harvest-aid Robot Scheduling Problem and the Implementation of Its Simulation Platfor

Manual harvesting contributes up to 60% of variable production costs for fresh market produce. Additionally, increasing worker shortage often lead to loss of production. Machine-aided harvesting can alleviate these problems by increasing picker efficiency. In strawberry harvesting, non-productive times include walking to the unloading station to deliver harvested strawberries and walking back to resume picking. A team of harvest-aid robots is being developed to carry full and empty trays between pickers and the unloading station. If properly scheduled, robots can reduce non-productive times.;Investigation and analysis on this Robot Scheduling Problem was performed to build a foundation for further research. Existing exact solution solvers were utilized and common scheduling policies were evaluated on the scheduling problem.;Manual harvesting for strawberries were observed to create an accurate model of the harvesting operation. Details and properties of the harvesting activity including picker decision policies, field covering pattern, pickers traveling velocities and action times were monitored and recorded to create a strawberry harvesting activities simulation platform.;Harvesting activities for manual and robot-aided harvesting were simulated. Dynamic scheduling policies such as "Shortest Processing Time" and "First Come First Served" were implemented and performance indicators like picker waiting time, and robot utilization rate were calculated. Simulation results showed furrow length is the primary factor affecting picker nonproductive time. Robot-aided harvesting also has a better improvement margin in fields with longer furrows. Differences among scheduling policies diminish when robot-picker ratio increases, and all policies generate same schedule when the ratio is high enough. Robot fleet size is the primary factor affecting picker waiting time. In a configuration where robot number increased from 2 to 4, the mean waiting time dropped 14%, and the maximum waiting time dropped 60%. The simulation platform will serve as an evaluation and prediction tool for further development of the Robot Scheduling Problem of the FRAIL-Bots project.