Project Overview
Commodities
Practices
Proposal abstract:
Fresh-market tomato production in Florida commonly relies on stake-and-string support systems, including the Florida weave, to reduce lodging, maintain plant structure, and support crop management throughout the season. However, tying and re-tying tomato rows remains a repetitive, labor-intensive task that must be completed within narrow management windows. Despite advances in mechanized planting, spraying, and other field operations, tomato tying remains largely manual because it requires crop-aware perception, safe interaction with living plants, flexible material handling, and reliable operation under variable field conditions. This project will develop and evaluate a low-cost robotic tying-assist system for sustainable fresh-market tomato production in Florida. First, it will develop a manipulation-oriented row perception module that identifies stakes, action reference points, approximate plant-base regions, canopy boundaries, and conservative safe work corridors. Second, it will develop a modular end-effector for twine dispensing, tension regulation, and repeatable tying or anchoring actions, with basic fault detection and recovery. Third, these modules will be integrated on a mobile platform and tested in representative tomato production settings. Expected outcomes include a prototype robotic tying-assist system, quantitative evaluation of perception and twine-manipulation performance, row-level comparison between robotic and manual tying operations, and practical guidance for future sustainable specialty-crop automation.
Project objectives from proposal:
Objective 1: Develop a manipulation-oriented row perception module for tomato tying that identifies or estimates key field elements needed for action, including stake locations, stake reference points, approximate plant-base or stem regions, canopy boundaries, and conservative safe work corridors.
Objective 2: Develop a low-cost robotic tying end-effector capable of twine dispensing, tension control, and repeatable tying or anchoring actions across neighboring stake segments, with detection and recovery strategies for common failure modes such as slipping, tangling, slack twine, or failed anchoring.
Objective 3: Integrate the perception module and tying end-effector onto a mobile platform and evaluate robotic versus manual tying at the row or task level in representative Florida tomato production environments, with metrics including tying success rate, placement consistency, task time, intervention frequency, and visible plant damage.