Computer-Aided Design and Optimization of a Multi-Level Fruit Catching System: Performance Evaluation of Fixed and Adjustable Designs for Soft Fruit Harvesting
Guardado en:
| Udgivet i: | ProQuest Dissertations and Theses (2025) |
|---|---|
| Hovedforfatter: | |
| Udgivet: |
ProQuest Dissertations & Theses
|
| Fag: | |
| Online adgang: | Citation/Abstract Full Text - PDF |
| Tags: |
Ingen Tags, Vær først til at tagge denne postø!
|
| Resumen: | Tree fruit harvesting is a labor-intensive operation, and effective Multi-Level Fruit Catching and Retrieval (MFCR) systems have been proposed to facilitate the mass harvesting of soft fruits using trunk shaking. However, excessive fruit damage from falling through the canopy poses a significant challenge, necessitating optimization of various MFCR design aspects. Our study introduces a novel computer-aided design approach employing geometry, simple mechanics, and simplified collision kinematics to refine MFCR system design parameters. This approach includes the development of two indices: one quantifying the interference of MFCR catching booms with tree branches and another representing accumulated fruit damage based on simplified collision kinematics. These indices are applied in a case study involving twenty digitized pear trees to determine the optimal number of MFCR layers, maximizing marketable fruit collection. This approach offers a foundation for optimizing fresh-fruit mass harvesters that use multi-level catching and retrieval systems. To further enhance the efficiency and adaptability of MFCR systems, we propose an adjustable MFCR system that dynamically modifies layer heights based on observed fruit distribution, offering a flexible alternative to traditional, fixed systems. Initially, we employ a genetic algorithm to benchmark a multi-level fruit-catching system, optimizing parameters such as the number of layers, layer heights, and the number of arms per layer. Using voxel-based GPU computing, we perform accelerated interference analysis across a broad range of harvester configurations. Subsequently, we validate the optimized MFCR against an adjustable design that modifies layer heights based on real-time visible fruit distribution, detected through simulations using YOLOv5-trained fruit detectors on synthetic image data. Additionally, we use Gurobi to optimize layer heights to minimize fruit dropping distances, thus reducing fruit damage. Performance comparisons indicate a 10% increase in marketable fruit collection with the adjustable MFCR, demonstrating its potential to improve efficiency, particularly when tree structure data is available. |
|---|---|
| ISBN: | 9798290614397 |
| Fuente: | ProQuest Dissertations & Theses Global |