About

Workshop Project Report 

1. Project Information

2. Rationale

In 2050, the global population is projected to be 9.7 billion. To meet the increasing demand for food, production must increase by an average of 60% in the next 30 years. Urbanization, land degradation and limited natural resources, and the aging and shortage of the agricultural workforce add more challenges to the existing problems of agricultural production under climate change. Technology innovations to enhance agricultural productivity is a high priority for many countries, including Taiwan. Smart agriculture (digital agriculture) is among the highly recommended innovations to help address the problem of farming labor shortage and enhance productivity and efficiency of agriculture and food systems. The scope of smart agriculture is wide, and this workshop focuses on the digital innovations of the agriculture sector, including crops, livestock and fisheries.

While implementing the digitalization of agriculture, countries face several challenges: big data ownership and cybersecurity; business model development and promotion; digital transformation, industrialization of the new technologies; supporting facilities and policies; overcoming the digital divide in rural infrastructure; and human resource and capacity development. To foster international cooperation and enhance smart agricultural development in the Indo-Asia-Pacific region, this workshop aims to invite experts from the region to (1) share country experiences and solutions in industrializing smart (digital) agriculture and discuss the challenges and strategies to implement smart agriculture; and (2) initiate an international consortium on smart agriculture. The consortium would provide a platform for continuous information exchanges, facilitate consensus building and setting regulations and guidelines for international application of digital agriculture technologies.

3. Objectives

• Exchange information on the progress, challenges, and strategies of smart agriculture development in the Global Cooperation and Training Framework (GCTF) partner countries.
• Identify challenges and strategies of internationalization and industrialization of smart agriculture, and identify the policies and facilities required to support and advance smart agriculture development.
• Initiate a GCTF smart agriculture consortium to facilitate partnership and setting guidelines for international application of smart agriculture technologies.

4. Themes

• Smart agriculture from research to development in the Asia-Pacific region – status and challenges exemplified by case reports from Indo-Asia-Pacific countries
• Business models of digital agriculture industry and their applications: cases and examples - digital twin, sharing economy of agricultural machinery, smart agriculture service industry, smart agriculture industrial ecology, 5G applications
• Digital transformation and supporting system – problems and solutions: agricultural data collection, utilization, sharing, ownership, regulations, cybersecurity; problems and responses; government roles, supporting policies and investments encouraging public and private partnership to reduce digital divide in rural infrastructure

5. Workshop highlights

The workshop organizers invited 11 speakers, 7 panelists (3 panelists also served as speakers) and 3 moderators from six countries (Australia, Israel, Japan, the Netherlands, Taiwan, USA) to present/ discuss smart agriculture development in their countries and views on the challenges and ways forward. The presentation materials will be shared on the GCTF workshop website for registered members to access. The workshop also produced an 80-minute video of online field trips which will be uploaded to the FFTC website for public access. Each workshop presentation by session is summarized below:

Session 1: Smart agriculture from research to development – status and challenges

Opportunities and challenges for artificial intelligence in smart agriculture

Speaker from AIFARMS, University of Illinois, USA, shared institutional research goals of AI in agriculture to tackle the two fundamental challenges of farm labor scarcity and heterogeneity of agricultural data across crops, animals, environmental conditions at  spatial and temporal scales. Six connected thrusts enabled by AI R&D were highlighted: (1) Autonomous farming: AI-driven low-cost robot teams enable reduced environmental harm and improved yields; (2) Labor optimization for livestock: Computer-vision for 24x7 livestock monitoring enables better animal welfare and lower labor requirements; (3) Environmental resilience: Sparse learning and machine learning (ML) for large hetero-data sets improve resource use efficiency, climate resilience; (4) Sil monitoring and health: Sensor networks & ML for large hetero-data sets enable scalable soil monitoring, nutrient flow prediction; (5) technology adoption and public policy: Farmer workshops/surveys and technology transfer to industry; and (6) Education and outreach: Digital agriculture courses, customer service on-ramp; K-12 outreach.

Development and strategies of smart agriculture in Taiwan

A speaker from the Council of Agriculture (COA), Taiwan shared the scope of Taiwan’s smart agriculture program and the government’s promotion efforts. The first phase started from 2017 with the vision of efficiency, safety, and low risk, aiming at helping shareholders to achieve a breakthrough in digital transformation. The COA prioritizes ten key agricultural industries and their related agri-technologies for country-wide promotion. The second phase will kick off in 2023 with the development goal of precision, collaboration, and popularization. The program will focus on industrialization of the innovative models and cost-effective machinery, automation and services. Public-private partnerships to form sound industrial agriculture ecosystems will be highlighted. 

Smart agriculture towards Society 5.0

The speaker from Hokkaido University, Japan introduced data-driven, automation and robotization of rice farming in Japan, and highlighted that smart agriculture is realized by the integration of cyber space and physical space. The speaker shared his research using 5G mobile network, AI and automation. Examples included: (1) a robot which can precisely apply fertilizers based on crop growth and soil conditions and detect and respond to diseases and insect damages at early stages; (2) robot boats for paddy field management; and (3) autonomous drones that not only spray pesticides but also collect crop growth information. The Japanese government has set the 2020 KPI for realization of agricultural robots which can move between field sites automatically and monitor the conditions of dispersed fields remotely. These technologies have been developed for rice farming in Japan. The next step is to develop technologies that allow one person to manage multiple robots and the smart robots for vegetable and fruit production.

Digitally-enabled innovation in industrialized agriculture

The speaker from the Commonwealth Scientific and Industrial Research Organization (CSIRO) in Australia introduced smart agriculture research and development in Australia. The key issue is how to support rapid adoption and innovation  by 2030. The roles of digital technology and innovations in smart agriculture are categorized into three types: automation; digitally enabled decision making; and system changes. System change is the most challenging as it consists of multi-components, highly mixed enterprises, nutrient/energy/waste circularity and intensification. To understand the system requires predictions of economic value propositions, landscape fit, infrastructure fit and interactions within the system such as disease risks. System change research, e.g., agri-system innovations modeling at paddock to landscape to continent scale are on-going at CSIRO.    

Innovation for improving farmers’ accessibility to data through digitalization

The speaker from the Ministry of Agriculture and Rural Development, Israel, introduced Israel’s agriculture, highlighting digital transformation and data accessibility. Israel is in a hotspot of climate change and has limited agricultural land and other resources, including fresh water. The Israeli government is actively promoting digitalization of agriculture and data accessibility to all farmers. Examples of the innovations include: (1) a support mechanism to facilitate financial aid for growing new crops; (2) an agricultural knowledge center that leverages Israeli research capabilities; (3) an investment fund to encourage farmer adoption of new technologies; (4) GIS-based web systems for streaming farmers’ work, natural resource utilization, a collaborative platform among various stakeholders, and a platform with integrated data and information; (5) a dashboard that allows data analysis on various layers of agriculture plots by crops, areas, irrigation and coverage etc.; (6) AgriMeteo – a weather forecasting application; and (7) the Agriculture 4.0 Paradigm.   

Dutch smart agriculture: turning data into dollars

The speaker from the Netherlands private sector shared the company’s mission and approaches to turning data to money. The Netherlands is an agricultural powerhouse and the 2^nd largest exporter of flowers, vegetables, and seeds. Business Intelligence, Internet of Things (IoT), AI, autonomous vehicles, and robotics are gradually being adopted in agriculture. By analyzing and using the data the company can create value for the Environment, Society (Consumers), and Businesses. (People, Planet, Profit). Data and automation can optimize and drive further development of the agricultural sector. Solutions should be incorporated within the existing infrastructure. The data can be turned into dollars and create value for the environment by better use of scarce resources and less waste. Consumers will benefit by receiving fresher, healthier and cheaper produce. Agri companies can increase output and quality and lower operational costs, thereby realizing a profitable business. Essential prerequisites to succeed include: (1) using alternative and new energy sources; (2) education: a continuing necessity analysis, interpreting, and managing the new realities; (3) cooperation:  teaming up with various key experts in their fields; and (4) supplying integrated chain solutions and turning data into dollars requires joining forces in horticulture alliances, sector associations, and international activities.

Session 2: Business models of digital agriculture industry and their applications

KSI's smart agriculture frontline experience sharing: forming a business ecosystem

The speaker from KSI, an information system engineer company in Taiwan, introduced the approach and timeline of applying a collaborative and integrated smart agriculture program at vegetable production farms in Yumei. In 2015-2019, KSI focused on information systems engineering of the production plan, material management, cultivation records, and cost management. The Digital Twin in 2022 leads KSI from digital to smart agriculture and helped to build up an AI model that is learned from experienced farmers’ behaviors and the model keeps on learning and evolving. To dive deeper into the smart agriculture and explore multiple possibilities, KSI partners with IoT technology providers, works with different university crop specialists, big data analysis companies, and field experts. Standardized cultivation processes, optimal cultivation, flexible locations with equal quality were achieved through the collaborative efforts, particularly with NTU, and a replicable success model can be applied to different crop production systems. Eventually a cultivation method (eg. Yumei) can be branded, and a digital intellectual property can be created and  traded.

AgriTalk’s smart agricultural solution and business model

The speaker from National Yang Ming Chiao Tung University, also the Founder of AgriTalk Technology Inc., Taiwan, emphasized the integration of user experience with IoT, AI and biotechnology to provide a green and smart solution. AgriTalk cooperates with Quanta Computer to develop sensors and control modules that assist farmers to monitor farms accurately in real-time, develops technology to predict soil microbiome distribution and disease and pest outbreaks; provides effective and specific bio-reagents for disease and pest control; and develops complete carbon credit solutions for farms. Three business models designed by AgriTalk based on customer demand, including farm automation solutions, farm AI solutions, and total solution export have been successfully promoted and have reached their customers in Taiwan, Thailand, Japan, and the Philippines for farming of tumeric, strawberry, tomato, asparagus, melon, and banana. 

Kubota's initiative on smart agriculture and future development

The speaker from Kubota, Japan shared Kubota’s innovations in smart rice farming systems. Innovations tackling farm labor scarcity and attracting the next generation of professional farmers include: (1) a Japanese-style precision farming system connected via smart farm machinery featuring a farm management and machine service support system that collects and makes use of work records and sensor data from connected harvesters and other farm machinery; and (2) ultra-labor-saving technology via automation and unmanned operation of farm machinery. The three levels of automation and unmanned operations defined by Japan’s Ministry of Agriculture, Forestry and Fisheries are L1: automatic driving while the operator is on the machine, L2: automatic and unmanned operations that are human monitored, L3: complete automation under remote monitoring. Kubota has developed various sized rice transplanters and tractors at L1 and L2 (“Agri-Robo” series), respectively. In the future, Kubota will attempt to expand smart agriculture not only for rice farming but also for vegetables and fruits, and will provide total solutions that will benefit the entire food value chain in Japan and Asia.

Business models of the digital agriculture industry and their applications

The speaker from Australian Centre for International Agricultural Research (ACIAR) addressed the need for supporting smallholders in smart agriculture development. Smallholder farmers who produce food for more than 80% of global populations will be caught up in the digital transformation and respond to the challenges by ‘stepping up’ or ‘stepping out’. Four types of digital agri-tech services that can transform smallholder agriculture were suggested, including: (1) digital advisory and extension that can provide information, data, and trainings to farmers; (2) digital financial services to spur the development of new or modify existing business models, applications, processes, and products; (3) digitalized farm tools using a range of technologies embedded within on-farm products; and (4) digital market linkages that connect farmers and intermediaries in the value chain and provide opportunities to standardize and streamline formal and informal business practices. Key messages are “smallholders respond rapidly to financial opportunity,” “business model innovation,” “knowledge brokering,” and “not all about digital”.

Smart pig farming inThe Netherlands

The speaker from Wageningen-UR Livestock Research, The Netherlands, pointed out the pressure of improving labor performance and automation to upgrade the piglet farming industry in The Netherlands and keeping it competitive for global markets. The speaker shared WUR’s recent work on the development of sensors for automatically measuring pig behavior and health. The sensor techniques, for example, are able to continuously monitor pig activity, posture, feeding, and drinking, and, from sound analysis, can measure e.g., coughing as an indicator of disease. These sensors can potentially support farmer management. Continuously improving these sensors and algorithms are on-going. Future effort is needed to use these technologies to develop early warning indicators of impaired resilience in pigs.

Session 3: Panel discussion- Digital transformation and supporting system

Topic 1: Digital transformation and supporting system – problems encountered and possible solutions

A panel of three members provided the following views to address the session topic.

Data-driven food Systems to sustainably nourish the world (Microsoft, USA)

Precision and regenerative agriculture has been shown to improve yield, reduce costs and ensure sustainability. High costs of manual data collection prevent farmers from using data-driven agriculture. Connectivity on farm and sparse sensor development are the key technical challenges. FarmBeats offered by Microsoft is a solution that provides farmers with access to the Microsoft Cloud and AI technologies, enabling data-driven decisions to help improve agricultural yield, lower overall costs, and reduce the environmental impact of agricultural production.

Taiwan government’s promotion and support for smart agriculture development  (TARI, Taiwan)

TARI envisions a “multi-ecosystems in an ecosystem” for smart agriculture. Starting in 2017 the nationwide promotion, Taiwan government provides parallel and serial projects to support the technology development, infrastructure development and industrialization of smart agriculture in Taiwan. Phase I (2017-2022) invests in R&D for technology breakthroughs and verification in the areas of smart production and digital services; Phase II (2023 – 2025) gives priority to agriculture IoT, cloud generation, digital transformation, business models and diffusion. Multidisciplinary, cross-sectoral cooperation and public-private partnership are highlighted.    

Digital transformation and supporting systems - Learnings from the Australian context (CSIRO, Australia)

• Need for an integrated solution
• Lesson learnt for success:
• Requirements: integrated data (data, analytics, visualization, decision); use-case driven; timely, accurate, scale, and fit-for-purpose; personal/business tailored; incentives for investments and/or data sharing
• Role of machine learning: in domain context, data science x simulation modelling
• Shifting focus: monitoring to prediction, farm to landscape and beyond, automation

Topic 2: A consortium for continuous information exchange, facilitating partnerships, and setting international guidelines for digital technology development

A pre-workshop survey on the proposed consortium was conducted among workshop participants. The moderator first presented the survey results followed by brief presentations of the panel of four members.   

Consortium survey results (conducted by FFTC and reported by the Moderator, NTU, Taiwan)

Among 374 invited workshop participants, 87 people from 29 countries provided their feedback. About 48% of the respondents are from governments, 38% from academia and 14% from the private sector. About 93% of the respondents think the consortium is necessary; 78% are interested to join the consortium and another 20% are considering joining. The majority of the respondents are fine with the proposed structure of the consortium. Some of them provided suggestions to improve the contents.

Data governance in smart agriculture (NTU, Taiwan)

The panelist indicated the complexity of data by listing the different sources and types of data and explained the flow from data to knowledge. Challenges of data governance in smart agriculture include data heterogeneity, interoperability, reliability, ownership, privacy, data and device security. Examples of facilitators of data governance were given and a conceptual framework of a consortium at the three dimensions of data flow, data governance, and stakeholders along the smart agriculture ecosystem was presented.      

Consortium – harness modern orchard system variabilities and seasonality

(Washington State University, USA)

The panelist expressed the need of a consortium and used the smart orchard as an example. Orchard systems are variable and seasonal, such as  the conventional ones, tall spindle, bi-axis and V system. For example, sprayer vehicles are commonly used in orchards, but the designs could be very different, and data are not compatible. An international consortium could help to bridge the technology gaps across the continents. In addition, a consortium could contribute to promote the standardization of data acquisition protocols to reduce data silos. Public-private partnerships and stakeholder education should be included.     

Consortium – a platform to integrate domain knowledge (Hokkaido University, Japan)

The panelist first presented the co-evaluation of AI and robots in agriculture that various agriculture sub-systems (domain) are mapped in “cyberspace” and AI and robots are co-evolved through digital twining to operate in the field. Domain knowledge is important, and the integration of these domains is critical to achieve smart agriculture. A consortium provides a platform that allows different people with different knowledge aspects and levels to communicate and work together.       

Getting greater value from agriculture data: Australia’s experience (CSIRO, Australia)

The panelist pointed out that even in Australia, one of the world leaders in smart agriculture, has many digital agriculture systems that are siloed. Proprietary data solutions that are good for immediate needs may not facilitate easy data sharing. Sharing of agricultural data more broadly across the agricultural supply chain can create greater benefits beyond immediate solutions. There is a lack of trusted solutions that make data discovery and sharing easy. The Australian government devotes great effort into platforms for data exchange, sharing and standardization. Consortium initiatives that serve the similar function are necessary.

6. Suggestions and conclusions

1. Adopt the low-cost semi-autonomous systems which can augment human efforts to increase yields with lower environmental impacts, better soil health, improved animal welfare. Novel machine learning techniques can analyze sparse and heterogeneous data at widely varying spatiotemporal scales that are pervasive in agricultural data sets.
2. Focus on realizing the use of innovative models and cost-effective automation as smart agriculture in Taiwan will enter the second stage in 2023, with “precision, collaboration, and popularization” as its development goal.
3. Adopt the system which enables robots to move autonomously between fields and a group of robots that work simultaneously in dispersed fields can significantly improve farm productivity and efficiency.
4. Learn the lessons from the Australian context in which agriculture innovation of multiple types and at multiple scales is accelerating. A major strategic target is system-level innovation that could lead to rapid step-change innovation. To successfully understand new systems and their interactions remains challenging.
5. Study and learn the growth of global agriculture production in the next 10 years which is attributed to yield improvements from increased input intensity, investment in production technology, and improved cultivation practices.
6. Optimize and drive further the development of the agricultural sector through data and automation. Solutions should be incorporated within the existing infrastructure.
7. Work with experts from different areas for companies to keep moving forward and the future is unlimited. IoT technology integration might be the end for an ICT company involved in agriculture.
8. Take customer demand into consideration and learn from AgriTalk’s experiences in their development of three different business models: farm automation solution, farm AI solution, and total solution export that have been successfully promoted in Taiwan and overseas.
9. Adopt important values for companies to deliver to customers working in smart agriculture including earning profits from agriculture by improving productivity, reducing heavy and high-risk work, reducing environmental stress, and revitalizing rural areas including the abandoned arable land.
10. Innovate digital agtech by focusing on providing financial opportunity, business model innovation, knowledge brokering and not-digital-at-all to unlock the potential of commercial smallholders to participate in smart agriculture.
11. Use new techniques and applications in pig farming to better manage animal health, welfare and economic results. Sensors to monitor health and early warning indicators of impaired resilience in pigs are important.
12. Support the private sector by including innovations that enable data-driven solutions to help farmers improve yields, lower costs, and reduce environmental impacts. Support from the public sector could include integrated solutions, national-wide promotion of smart agriculture development with clear policies enabling PPP, formation of ecosystems (and sub-systems), incentives for investment, data sharing, and shifting the focus from monitoring to prediction, farm to landscape and beyond, and automation.
13. Consider forming an international consortium on smart agriculture as it is necessary to serve as a platform for workshop participants to continue communication, share information and do networking in accelerating digital transformation of agriculture in the developing world. All the panelists are positive about the consortium and shared their views on its possible scope. However, the areas of work should be further identified and focused.

7. Outcomes

More than 370 people from 48 countries across 4 continents, registered for the workshop, who were invited by the GCTF partner agencies. The top-ten countries were Taiwan (108), Japan (27), Australia (23), Honduras (22), Paraguay (14), Ghana (13), Vietnam (12), Sri Lanka (9), Philippines (8). Most participants were from government and public R&D institutions. The evaluation form was sent to participants the day after the workshop and 38 people provided their feedback. Most respondents were satisfied with the workshop in the aspects of content, relevance, and logistics. However, the workshop time (0900-1300, GMT+8) did not fit well for Africans and Latin Americans who made up than 30% of the registrants. It is also unfortunate that the workshop materials are restricted to GCTF members only according to the GCTF policy. The workshop was regarded as successful in terms of planning, coordination, and execution.

GCTF website: https://www.gctf.tw/en/

Workshop background at

https://km.fftc.org.tw/workshop/7

The Smart Agriculture in Taiwan - online field trip can be also watched at FFTC-KM YouTube with Playlist 

https://youtube.com/playlist?list=PLWLBOqIsQNRoI3C65JbAZbdqoofkAK2N4