A-PHP – Autonomous Pruning Harvesting Payload
- Call title: PR-ERDF 21-27 PRI-1-2 Support for the early-stage growth of innovative start-ups
- Project title: A-PHP – Autonomous Pruning Harvesting Payload
- Application number: FD206620
- Project code: 1003206620
- CUP Code: J19J24000910006
- Investment cost: €299,023.81
- Granted contribution: €149,511.91
The A-PHP (Autonomous Pruning Harvesting Payload) project, funded under the Finpiemonte call for support for innovative start-ups, is developing a robotic payload dedicated to automated grape harvesting, designed to be mounted on ground or aerial agricultural platforms (UGV/UAV).
Italian and European viticulture currently face crucial challenges: a shortage of skilled labor, rising production costs, and pressure toward more sustainable practices. The A-PHP project addresses these needs by automating harvesting, one of the most expensive and labor-intensive operations in the sector.
Technological objectives
The project plans to develop and integrate several advanced subsystems:
Computer Vision System
Industrial cameras and deep learning algorithms to recognize bunches, vines, and leaves in real time, precisely identifying ripe bunches ready for harvest under varying environmental conditions.
Planning and Control Algorithms
Proprietary AI software to calculate optimal trajectories and harvesting modes, ensuring operational speed and the integrity of the harvested bunches.
Multiple, Independent Robotic End-effectors
Robotic tools equipped with a peduncle cutting system and variable-geometry collection cones, capable of picking bunches with a delicacy comparable to manual harvesting, operating independently on three axes.
Hardware/Software Integration
A system compatible with various existing agricultural vehicle platforms, requiring no structural modifications to the vineyards.
Project activities
Requirements Analysis
Definition of functional specifications based on the operational needs of wineries.
Vision System Development
Training and validation of neural networks for the recognition and localization of grape bunches on the vine.
End-effector Design and Construction
Building prototypes for the delicate and precise harvesting of bunches, with a conveyance system to the storage container.
Laboratory Integration and Testing
Validation of subsystems and communication interfaces in a controlled environment.
Vineyard Experimentation
Field tests to evaluate the reliability and performance of the payload under real harvesting conditions.
Optimization and Documentation
System refinement and preparation of technical documentation in preparation for future CE certification.
Innovation
The A-PHP system distinguishes itself from traditional harvesters—which operate by shaking the entire plant—through its ability to harvest bunches selectively and precisely, minimizing damage to the plant and the product. Furthermore, it operates in vineyards not structured for mechanization, without requiring costly restructuring, and is compatible with existing agricultural platforms.
Expected impacts
Economic
Reduction in harvesting costs and increased productivity for wineries of all sizes.
Social
Response to seasonal labor shortages and a reduction in strenuous and repetitive manual labor.
Technological
Development of advanced skills in robotics and AI applied to precision harvesting, with positive repercussions for the Piedmont innovation ecosystem.
Environmental
More efficient and selective harvesting, with less impact on the plant and reduced waste compared to traditional mechanical solutions.
Conclusion
The A-PHP project, supported by Finpiemonte, represents a concrete step toward the intelligent automation of harvesting. By combining robotics, computer vision, and artificial intelligence, it aims to offer the viticulture sector an innovative and scalable tool, contributing to the establishment of a sustainable and competitive Agriculture 4.0.