Refill Scheduling for Agricultural Robots and Other Vehicles

Rob Fitch

Refill scheduling is the problem of deciding when a robot or other agricultural vehicle should pause in its work to replenish a resource, such as herbicide or fuel. This problem is commonly solved in broadcast spraying, for example, by simply running the spray tank dry and then refilling it.

This strategy actually leads to lost time in traveling to the refill location, and we can show that these time losses can be significant. When multiple machines must queue at a refill location, the problem is made worse.

In this talk, Rob will explain the theoretical difficulty of this problem and give examples from robotic spot-spraying and broadcast spraying to illustrate the potential time losses. He will present an optimisation approach that chooses optimal refill times to minimise travel distance and queuing time. These results apply to agricultural robots, human-driven spray rigs, and any other machine that must refill or empty some material at a fixed location during the course of its work.

Rob will conclude the talk by tying these results into the larger research program in agricultural robotics, including novel machine learning methods for fruit/vegetable detection that support selective harvesting.

Rob Fitch is Associate Professor at University of Technology Sydney.He was previously a Senior Research Fellow with the Australian Centre for Field Robotics (ACFR) at The University of Sydney where he retains an honorary position. He is a leading research scientist in the area of autonomous field robotics. He is interested in systems of outdoor robots and their application to key problems in agriculture and environmental monitoring.

Robert received his PhD in computer science from Dartmouth (USA). He has led research in planning and collaborative decision-making for both ground and aerial robots in a variety of government and industry sponsored projects including those in broad-acre agriculture, horticulture, bird tracking, and commercial aviation.