NRI: INT: COLLAB: Distributed co-Robots for Strawberry Harvesting

NRI：INT：COLLAB：用于草莓采摘的分布式协作机器人

• 批准号: 1924640
• 负责人: Manoj Karkee
• 金额: $ 30.29万
• 依托单位: Washington State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-09-01 至 2024-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1924640&HistoricalAwards=false
• 关键词: NRI; INT; COLLAB; Distributed; co

Recently, strawberry production in the United States is decreasing due to labor shortages and the ever-increasing cost of labor, particularly for harvesting. A high cost of production could put strawberry-growers in the U.S. at a disadvantage in comparison to countries whose cost of labor is much less. In order to make strawberry production in the U.S. competitive with the rest of the world, production costs need to be reduced. This project aims to investigate developing a small, rugged, and cost-effective co-robot system that can be used by growers, which will ultimately lower the harvesting costs and increase profitability of strawberry production. Additionally, the research results can benefit other high value specialty crops, such as tomatoes and blueberries, as they are facing similar labor shortage issues. This multi-disciplinary research will enrich course curricula which benefits a large number of students at three participating universities, and will provide research and development experiences to students of underrepresented and minority backgrounds. The outreach component of the project will educate K-12 students and the general public about robotics in agriculture via conferences, workshops, field days, and other learning engagement activities. Additionally, a graphic user interface will be developed, which is adaptable as an assistive device to benefit users with disabilities. The project includes three primary research thrusts. The first thrust will investigate a decentralized, scalable row allocation algorithm for robots to achieve a consensus about which rows to harvest, in order to achieve an overall minimum time in harvesting. The algorithm is expected to be scalable, efficient, fast, non-conflicting, and simple. The second thrust will develop an integrated multi-exposure fusion, curvature analysis and hierarchical image processing algorithm. Supported by the end-to-end deep learning technique, the algorithm will detect strawberries, stems, and vines in field environments with a high accuracy. The third research thrust involves a study of a parallel delta robot-arm based manipulation system for picking and transporting harvested fruits. The associated amount of movements will be minimized. These three thrusts will be complemented by many other engineering tasks and a detailed cost analysis.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

最近，由于劳动力短缺和劳动力成本（特别是采摘成本）不断增加，美国草莓产量正在下降。与劳动力成本低得多的国家相比，高生产成本可能使美国的草莓种植者处于不利地位。为了使美国的草莓生产与世界其他国家相比具有竞争力，需要降低生产成本。该项目旨在研究开发一种可供种植者使用的小型、坚固且经济高效的协作机器人系统，最终将降低草莓采收成本并提高草莓生产的盈利能力。此外，研究结果还可以使其他高价值特种作物受益，例如西红柿和蓝莓，因为它们也面临着类似的劳动力短缺问题。这项多学科研究将丰富课程课程，使三所参与大学的大量学生受益，并将为代表性不足和少数族裔背景的学生提供研究和开发经验。该项目的外展部分将通过会议、研讨会、实地考察日和其他学习参与活动，向 K-12 学生和公众介绍农业机器人技术。此外，还将开发图形用户界面，该界面可作为辅助设备，为残疾用户带来好处。该项目包括三个主要研究方向。第一个重点将研究一种分散的、可扩展的行分配算法，以便机器人就收割哪些行达成共识，从而实现收割的总体最短时间。该算法预计具有可扩展性、高效、快速、无冲突且简单。第二个重点将开发集成的多重曝光融合、曲率分析和分层图像处理算法。在端到端深度学习技术的支持下，该算法将高精度地检测田间环境中的草莓、茎和藤蔓。第三个研究重点涉及基于平行三角机器人臂的操纵系统的研究，用于采摘和运输收获的水果。相关的移动量将被最小化。这三个目标将得到许多其他工程任务和详细成本分析的补充。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Real-time detection based on YOLOv5s with a modified neck for picking region of ridge strawberry.

基于YOLOv5s改进颈部的脊草莓采摘区域实时检测。

• 发表时间: 2022-07
• 期刊: USA
• 作者: He; Z.
• 通讯作者: Z.

Detecting and Localizing Strawberry Centers for Robotic Harvesting in Field Environment

检测和定位田间环境中机器人采摘草莓中心

• DOI: 10.1016/j.ifacol.2022.11.110
• 发表时间: 2022-01
• 期刊: IFAC-PapersOnLine
• 作者: He, Zixuan;Karkee, Manoj;Zhang, Qin
• 通讯作者: Zhang, Qin