Stavros Vougioukas, University of California, Davis, USA: Generating 3D Fruit Maps for Model-Based Assessment of Robotic Fruit Harvesting Efficiency
Engineers working in fruit harvest mechanization lack the modeling tools that would enable them to investigate the coupled effects of orchard layout, tree structure, fruit distribution, and harvester design on fruit picking productivity. Instead, engineers rely on a prototype-building and field-testing cycle, which is slow, costly, and cannot explore the above-mentioned interactions.
Our goal is to build and validate an open-source design tool that integrates mechanical simulation with validated models of fruit-bearing orchard trees, in order to evaluate the fruit picking productivity of selective fruit picking systems. As a first step in this direction, a novel method is presented, which utilizes high-frequency radio signals and trilateration to measure the locations of thousands of fruits at a rate of approximately one fruit per second. Location data were collected in numerous California pear and cling peach orchards for thousands of fruits. Based on these data, fruit maps and statistics—such as height distribution and horizontal distance distribution from the row centers—were computed.
Stavros Vougioukas joined the Department of Biological and Agricultural Engineering at the University of California, Davis in 2012. He is an Assistant Professor and works in the area of mechanization and automation for specialty crops. He has a Ph.D. in Robotics and Automation from Electrical, Computers and Systems Engineering, Rensselaer Polytechnic Institute, Troy, NY. His current projects include: tree and fruit digitization and geometric modeling; model-based design and evaluation of fruit picking actuators; agricultural labor activities modeling for human-robot collaboration; robotic harvest-aiding carts for in-field fruit transport; automated high-throughput strawberry calyx removal.