Location: Sugarbeet and Bean Research
Project Number: 5050-30600-001-000-D
Project Type: In-House Appropriated
Start Date: May 6, 2025
End Date: May 5, 2030
Objective:
Objective 1: Develop an automated, AI-based sensing system for real-time, precision assessment of crop yield, quality/maturity, and optimum harvest time.
Subobjective 1.A: Determine the optimum sensor configurations and wavelengths for fruit size, yield and maturity assessments.
Subobjective 1.B: Design, assemble and test an autonomous mobile platform with multi-sensors for automatic acquisition of site-specific fruit maturity and yield information.
Objective 2: Develop a new generation of harvesting robots with substantially enhanced perception and manipulation functions to enable commercial full and selective harvesting of apple fruits.
Subobjective 2.A: Design and build a new, improved version of the perception module with greatly enhanced capabilities for accurate localization of fruits and branches under challenging tree structures and adverse lighting conditions.
Subobjective 2.B: Design and build a new version of the dual-arm manipulators with greatly improved planning and coordination algorithms for faster, more efficient detachment of fruits.
Subobjective 2.C: Incorporate a fruit ripeness evaluation function with the robot’s planning and control algorithm for selective harvesting of apples.
Approach:
Apples and other tree fruits are still harvested manually. Harvesting labor is the single largest cost in apple production and has become a top concern for apple growers due to the rising cost and growing shortage of labor. Site- or tree-specific information on fruit maturity or quality and yield pre- and at harvest is needed for precision orchard management, optimum harvest timing, selective harvesting by robotic harvesters and thus postharvest quality and marketability. This research is therefore aimed at developing innovative technologies for precision assessment of fruit size, yield and maturity pre- and at harvest and for automated harvesting of apples to help growers alleviate labor shortage and cost and enhance postharvest product quality and marketability. Cost-effective AI-based sensing configurations will be designed for accurate fruit size, yield and maturity assessment in orchard. An autonomous mobile platform with the new multi-sensor configurations will be built for real-time acquisition of site-specific fruit yield and maturity information in orchard to help growers make optimum orchard management and harvest timing decisions. Moreover, built on our recent progress in robotic harvesting, we will develop a new generation of the robotic harvesting system with substantially improved perception and manipulation functions for fruit detection, localization and picking (full and selective) under challenging tree structures with fruit occlusions by leaves and branches. The new technologies developed from the research will enable growers to reduce the reliance on manual labor in fruit harvesting and make better, informed orchard management and harvesting decisions to enhance postharvest product quality and marketability.