NSF Convergence Accelerator Track J: Data-driven Agriculture to Bridge Small Farms to Regional Food Supply Chains (L02619644)

全球气候变化和流行病暴露了全球化粮食系统的脆弱性，加剧了粮食不安全状况，特别是对于已经无法获得安全和负担得起的粮食和营养的多样化和服务不足的社区而言。由于需要有弹性的本地食品供应，美国重新把重点放在国内采购食品上。然而，物流和市场知识障碍限制了生产性地方粮食系统的生存能力。多个科学研究领域和人工智能、机器学习等现代技术创新的融合，可以通过扩大小农获取市场洞察的渠道，提高供需效率。该项目将使区域食品生产者能够了解其农场的特色作物品种和食用动物的经济价值，并将其与机构销售（如零售商、食品中心、分销商、杂货店、餐馆、医院、学校或学院）的市场需求进行比较。终端平台的数据驱动市场和财务洞察力将提高区域食品生产商与机构买家签订采购合同的能力，支持当地农场将其产品推向商店货架、餐厅餐桌和自助餐厅。应对区域粮食系统的挑战将对小农和当地企业的经济生计产生广泛的社会影响，并对增加安全和营养的当地食物的供应产生广泛的社会影响，这些食物将支持代谢健康，特别是对弱势社区。此外，加强对销售渠道的了解将减少粮食损失，增强作物多样性，从而为农民实践混合耕作和作物多样化等气候友好型可再生农业技术创造收入来源。最终，该项目通过对食品和营养安全的关注，促进了美国人口的健康和繁荣，以及环境管理。该项目评估用户需求，设计一个可扩展的技术平台，为小农提供市场洞察。主要研究目标是：(a)了解小农在向机构市场销售时遇到的知识障碍；(b)将多种科学学科的使用启发研究和新颖的数据驱动技术结合起来，为支持小农获取相关市场信息的软件平台开发概念设计。研究方法包括：(a)与小农和其他利益相关者一起发现他们面临的障碍；(b)市场分析；(c)数据收集，有助于概念设计和平台中计算模型的数据输入。数据收集包括：(i)库存评估和与服务不足的试点地区的机构买家面谈，以确定当地对食品的需求；（ii）通过机器人、遥感、卫星数据或无人机在农场收集的产品级数据，包括现有数据集和从种植者收集的数据；（iii）评估低成本有效的农场预防控制和微生物风险检测，以分析食品安全经济风险模型，以支持生产决策。作为将研究发现转化为市场影响的核心，该项目将确定小农在了解机构市场需求并向机构买家销售产品时遇到的障碍。通过确定知识差距导致供需效率低下的地方，它还将扩展对跨科学领域数据的理解，这些数据可以整合为商业决策提供信息，利用人工智能（AI）和机器学习（ML）技术，如计算机视觉，为农产品定价，并创建预测模型，以预测未来的粮食需求和定价。这项工作将推动小生产者数据驱动农业领域的发展，支持他们的生计、当地经济增长和粮食安全。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Global climate change and pandemic have exposed vulnerabilities in the globalized food system, exacerbating food insecurity, especially for diverse and underserved communities who already experience disproportionate access to safe and affordable food and nutrition. The need for resilient local food supply has refocused efforts on domestic sourcing of food in the United States. Yet, logistical and market knowledge barriers limit the viability of productive local food systems. The convergence of multiple scientific research fields and modern technological innovations such as artificial intelligence and machine learning can improve supply and demand efficiencies by extending small farmers’ access to market insights. This project will empower regional food producers to understand the economic value of the specialty crop assortment and food animals on their farms in comparison to market demand for institutional sales (e.g., retailers, food hubs, distributors, grocers, restaurants, hospitals, schools or colleges). The end-platform’s data-driven market and financial insights will enhance regional food producers’ abilities to obtain procurement contracts with institutional buyers, supporting local farms to bring their products to store shelves, restaurant tables, and cafeterias. Addressing the challenges of regional food systems will have broad societal implications for the economic livelihoods of small farmers and local businesses, and for the increased availability of safe and nutritious local food that will support metabolic health, particularly for disadvantaged communities. Furthermore, enhanced knowledge of sales channels will reduce food losses and enhance crop diversity, thus creating income streams for farmers practicing climate-friendly regenerative agricultural techniques such as mixed farming and crop diversification. Ultimately, this project advances the health and prosperity of the United States’ population, as well as environmental stewardship, through its focus on food and nutrition security. This project assesses user needs to design a scalable technology platform that provides market insights to small farmers. The primary research objectives are: (a) to understand the knowledge barriers that small farmers experience to sell to institutional markets; and (b) to converge use-inspired research of multiple scientific disciplines and novel data-driven techniques to develop the conceptual design for a software platform that would support small farmers to access the relevant market information. The research methods include: (a) user discovery with small farmers and other stakeholders about the barriers that they face; (b) market analysis; (c) data collection that contributes to the conceptual design and data feeds for computational models in the platform. Data collection includes: (i) inventory assessments and interviews with institutional buyers in the underserved pilot regions to identify local demand for food products; (ii) product-level data collected on-farm via robotics, remote sensing, satellite data or drone, including both existing datasets and data collected from growers; (iii) assessment of low-cost validated on-farm preventative controls and detection of microbial risk for analysis of food safety economic risk models to support production decisions. Central to the translation of research discovery to market impact, this project will identify the barriers that small farmers experience to understand institutional market demand and sell to institutional buyers. By identifying where gaps in knowledge contribute to supply and demand inefficiencies, it will also extend understanding of the data across scientific fields that could be integrated to inform business decisions, leveraging artificial intelligence (AI) and machine learning (ML) techniques, such as computer vision, to price farm products and create predictive models to anticipate future food demand and pricing. This work will advance the field for data-driven agriculture for small producers, supporting their livelihoods and local economic growth and food security.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.