Download PDF Summer Bridge on Smart Agriculture June 15, 2022 Volume 52 Issue 2 People everywhere rely on agriculture in one form or another – for food, animal feed, fiber, and other necessities. The summer 2022 articles describe precision indoor farming and alternative protein food systems, advances in food processing, genome editing, digitalization, sustainable and regenerative agriculture, the role of a circular economy, and the important role of policy. The Digitalization of Production Agriculture Tuesday, June 14, 2022 Author: John K. Schueller and John F. Reid System-oriented precision agriculture production leverages sensors, machine learning, and cloud computing to enhance productivity. Agricultural productivity has expanded for over a century because of successful replacement of human and animal power with mechanization combined with advances in fertilization, agricultural chemicals, and genetics. These developments freed workforces in developed countries to focus on building and advancing their societies. A second wave of productivity expansion has occurred since the 1980s that can be associated with the impact of digital technologies following the trends of Moore’s law. The result is system-oriented precision agriculture production solutions that leverage sensors, machine learning, and cloud computing to achieve productive outcomes. Continued expansion of digital technologies and widespread connectivity are important enablers for the development of future production systems for food, feed, fiber, and fuel that are more economically, environmentally, and socially sustainable. Fully integrated precision agricultural production systems link a network of processes, organizations, and stakeholder relationships and enable higher levels of optimization. Digitalization can improve agricultural production to meet the needs of increasing populations and changing diets. It also provides tools to adopt and evolve new production practices that address degradations (e.g., salinization, desertification, erosion) and facilitate progress toward sustainability objectives. Early Digital Agriculture: Promises and Challenges Digital agriculture began in the 1980s as machinery was endowed with electronics, embedded computing, software, and other digital technologies. Agricultural adoption of digital technology has typically closely followed similar enabling technologies in automotive and other industry verticals. Initially electronics and controls were incorporated in tractors and self-propelled vehicles. Automatic control systems ensured that vehicle operations were optimized according to control setpoints. Cooperation between European and North American engineers led to the ISO 11783 standard, which defined communications between tractors and attached implements over a CAN bus (Oksanen and Auernhammer 2021). But lacking external sensing, communications, and other off-board technologies, agricultural vehicles were “islands of automation” limited to specific machines. Automatic guidance supports reduced operator fatigue, consistent job performance, 24-hour operational capabilities, and increased field operation accuracy and productivity. The ability to make the most of mechanization had a practical limit. Larger, more powerful equipment and fewer field passes increased productivity with a reduced agricultural workforce, but fields’ varying topography and soil characteristics caused nonuniform yield outcomes. The concept of precision agriculture was founded on the principle of using advances in mechanization to address spatial variability. Pioneering research and commercialization made it possible to spatially sense and map instantaneous crop harvest yields on the go, and that remains a core crop sensing solution for modern combine harvesters (Myers 1996; Searcy et al. 1989). The integration of crop yield sensing with the location of the machine in the field via the Global Positioning System (GPS) provided farmers with basic geospatial information to make (limited) strategic or tactical decisions (Schueller 1992), such as whether to improve field drainage or spatially vary fertilization. Some early solutions were complex to set up and manage, though, frustrating farmers with the planning and preparation required before productive work could begin. And although GPS promised future benefits, its accuracy was initially adequate only for large harvesting machines. Early precision agriculture created a vision of tighter integration of machine systems to provide better agronomic outcomes, but the reality was challenging because of missing or erroneous data, unknown accuracies, weather, and complex physical/chemical/biological plant growth processes. Value propositions struggled to move beyond these challenges until automatic guidance evolved to enhance machine productivity and reset expectations. Enhanced Accuracy through Automatic Guidance Agricultural vehicle guidance was a game-changing innovation that offset some early precision agriculture challenges. While vehicle guidance solutions had been proposed or demonstrated since the 1920s (Wilrodt 1924), commercial realization required vehicle electronics technology combined with electro-hydraulic steering controls that became viable only in the 1990s. Guidance required position sensing enabled by public availability of the GPS network and subsequent advances that increased precision and accuracy to the centimeter levels necessary to meet various agricultural application requirements. Automatic guidance unlocked significant value that varied with the machine system, operator, and crop production system. Its benefits included less operator fatigue, consistent job performance, operational capabilities day or night, increased field operation accuracy, higher productivity (through higher-speed operations), and increased crop density through accuracy in planting patterns. Guidance requirements depend on the specific agricultural field operations and vehicle system characteristics. Meter-level sensing accuracies are suitable for yield sensing in harvesting operations, which have a wide swath. Decimeter-level accuracies through dual-frequency GPS are needed for tillage and basic spraying systems, and they provide information required for weed control or chemical placement with minimal overlap between machine passes. And centimeter-level accuracies, provided by real-time kinematic positioning–GPS systems, enable control of vehicle guidance proximate to rows of plants. As GPS accuracy improved, automatic guidance displaced manual steering with closed-loop control of the vehicle to desired straight lines or curved paths (figure 1). Early solutions in the market were options or aftermarket retrofits to vehicles, but over time factory-integrated solutions emerged as guidance and vehicle control became mainstream. GPS expanded to broader Global Navigation Satellite System (GNSS) solutions that leverage multiple constellations (e.g., Galileo from the European Union and BeiDou from China) to improve positioning solutions. Agricultural guidance requires understanding of vehicle system characteristics (for prime movers and implements), vehicle pose with respect to the terrain, and soil surface properties. And the sensing and control need to control the working implements to a precise location proximate to crops, sometimes traversing between planted crop rows. There remain some guidance use cases requiring a solution to sense and follow crop rows or to operate relative to a given crop. Mechanical feelers are a basic solution to sense and follow sturdy, established crops. And machine vision and lidar solutions have succeeded in sensing guidance cues for emerged crop applications and crop edge during harvesting operations. Examples of the advanced farming practices that have emerged include the capacity to follow precise paths from previous operations, field-level path planning to optimize path traverses, documentation of agricultural practices at the row or plant level, ability of vehicle systems to operate relative to infrastructure (e.g., drainage and irrigation systems), tractor-implement management, compaction management, and other forms of protection for cropping systems and plants. Machine Precision and Efficiency through Electrification During the past decade electrification has enabled greater precision and accuracy, for example in planting. In traditional planting systems, seeds are separated (singulated) from hoppers through mechanical or vacuum means and then dropped to the soil trench through gravity-feed seed tubes. The seed interaction with the tubes effectively limits planting speeds to less than 3 m/s for consistent seed spacing. Electrification enables independent closed-loop control to separate and transport individual seeds to the ground (Garner et al. 2013). Electric drives on independent row units not only precisely place seeds at speeds approaching 5 m/s with precision control but also solve challenges in mechanical-drive seeding systems to produce fields where every plant emerges at the desired location and depth. And these systems enable potential new forms of crop production such as grid-based farming methods (Kremmer et al. 2021). High-voltage electrification solutions in transportation verticals are beginning to be commercialized in production agriculture, improving the performance of vehicle drivetrains for traction and task automation. Increasingly electrified machine forms will create additional opportunities for production system efficiencies through precision control and systems-level power efficiency gains. And systems electrification will be part of future solutions that farmers can leverage in the transition toward alternative energy sources to support decarbonization (e.g., Gavioli 2021, p. 6). FIGURE 2 Late 20th century concept of precision agriculture from Hermann Auernhammer. Wireless communications and the internet have replaced datacards and rotary phones, and drones supplement the remote sensing of satellites and aircraft, but the agricultural activities remain similar. NH4 = ammonium. Reprinted with permission from Auernhammer and Schueller (1999). Digital Expansion of Precision Agriculture Digital technologies (machine-to-X connectivity, cloud, and mobile) are a hinge point for modern precision agriculture, enabling new ways to enhance production systems through the flow and management of information resources. Connectivity Information from tractors and mobile machines used for planting, crop protection, nourishment, and harvesting operations must be uploaded to management computers to support operations. Similarly, settings and actuator commands must be downloaded to the machines to drive field outcomes. Multiple vehicles working collaboratively in the same field must communicate to optimize performance. Connectivity technologies couple information transfer for machine-to-machine interactions and off-vehicle storage of agronomic and vehicle data to support job steps in crop production system performance. Connected agricultural vehicles with bidirectional information flow are mobile sense-and-control platforms. Connectivity also enables the use of off-vehicle resources for precision management (figure 2). Soil, weather, crop, and pest data can be used for crop management decisions closer to real-time operations. Lack of connectivity in some rural areas will affect realization of digital agriculture. Some producers are overcoming this barrier by installing private connectivity infrastructure,[1] and others have conceived of field-edge networks for local data transfer and storage. Without connectivity, farmers are limited to manual data transfer and other inefficient methods of information exchange. In the past decade, cloud computing capabilities have emerged and matured for agriculture, enhancing data management and storage and opening opportunities for value-added services. For example, the cloud has enabled the integration of crop modeling with remote sensing, resulting in predictive information for in-season crop management and yield. Furthermore, cloud-based management enables insights from vehicle use to measure operational effectiveness, including the ability to infer information on operator performance (Pfeiffer and Blank 2014). Drone Use Low-cost drones can be deployed for precision data acquisition, quickly traversing fields, collecting images, or performing operations without disturbing soils or crops. They can be flown exactly when and where desired to efficiently collect information for crop management or vehicle control. Field scouting via drones can map crop growth and identify pest infestations with high precision. Drones can scale to apply pesticides or fertilizers, either in response to sensed infestations or based on previously generated application maps, without the need to traverse fields with heavy equipment. Drones can be integrated with terrestrial agricultural vehicles performing field operations (Sugumaran 2017). Launched from a vehicle, a drone can take an appropriate aerial vantage point to assess conditions ahead of the vehicle or to inspect vehicle task performance. But given that farmers are overloaded with demanding and varying tasks, agricultural drones can achieve maximum utility only if they are easy to operate, seamlessly integrate into the production process, and quickly translate collected data for management actions. From Automation to Autonomy: Opportunities and Challenges Many of the technical elements that have enabled precision agricultural production systems also are critical for autonomous machine deployment (Reid et al. 2016), including precise positioning, path planning, connectivity, cloud, and drones. Opportunities Likely advantages are clear; for example: Autonomous systems may be part of the solution for agriculture’s varying labor requirements during the crop production cycle by augmenting the labor force. Autonomous vehicles can consistently perform over longer durations (day and night) without performance degradation. And autonomy may enable some operations to be performed without interruption and/or simultaneously. For instance, autonomy allows combine harvesters to unload grain on the go into driverless carts to improve in-field transport logistics with reduced labor. Autonomy opens possibilities for innovative machine forms that may become part of the resources used in determining total factor productivity (Reid 2011). Small autonomous machines are available for executing jobs with less soil compaction and other benefits, like operating in growing crops. One example is a robotic interrow solution that traverses standing crop and enables new management practices such as precision in-season nutrient application and cover crop planting ahead of harvest (figure 3). Challenges Despite nearly unlimited possibilities, there are significant challenges for the successful implementation of autonomy, starting with the need to identify use cases with reasonable paybacks. Agricultural seasonality and the associated importance of timeliness in crop production increase the hourly value of autonomy, but many tasks have limits to the hours of operation in a crop year for a specific production step. For commercialization success, autonomous agricultural vehicles must find tasks that provide enough economic benefit through labor savings, higher job quality, or removal of humans from dangerous tasks. The total job requirements must also be considered. Autonomous operation is more than machine systems executing the task. There are also setup, planning, logistics, documentation, and management of the autonomous operation. More research needs to approach autonomy from the perspective of the jobs to be done (Christensen and Raynor 2004). Design and programming must consider how the vehicle systems get to desired fields to execute production tasks and the farmer’s role in autonomous mission execution. In addition, perception sensing is needed for vehicle situation awareness. Such sensing is maturing, but requires further development to outperform an operator’s capabilities to sense proper machine operation and variable field conditions. And sensing is often difficult through impairments in the normal operation of the machine (e.g., from precipitation or dust). Artificial intelligence, such as computer vision/machine learning approaches, may provide value in precision agriculture applications. A challenge is the acquisition of significant image data representing the extremely wide variety of crops, production situations, and soil types. Summary Agriculture mechanization is advancing toward digital precision agriculture. Electronics and digital technologies commercially available in other verticals have been adopted and resulted in higher systems-level productivity of people, vehicles, fields, and farms. The precision and accuracy of automatic guidance create opportunities for further agronomic and management advances and contributions from agricultural vehicle systems. Digitalization also creates value through data that can be leveraged beyond the vehicle system and the farm to provide additional commercial and societal benefit. Societal investments in assets such as GPS and the internet have accelerated the transformation of agriculture. The technological advances in agriculture are timely to address challenges of increasing demands on global food supply. Digital precision agriculture provides new tools that better document production practices and enable the optimization of production systems toward decarbonization, regenerative farming approaches, and reduced impacts on climate change. Precision agriculture technologies can also shed light on new methods and increase efficiencies in production agriculture practices. References Auernhammer H, Schueller JK. 1999. Part 3.2 Precision farming. In: CIGR Handbook of Agricultural Engineering, Vol III: Plant Production Engineering, eds Stout BA, Cheze B. St. Joseph MI: American Society of Agricultural Engineers. Christensen C, Raynor M. 2003. The Innovator’s Solution: Creating and Sustaining Successful Growth. Cambridge MA: Harvard University Press. Garner E, Friestad ME, Mariman NA, Rylander DJ, Thiemke DB. 2013. Seeding machine with seed delivery system. US Patent 20130036956. Gavioli G. 2021. A sustainable mechanization for the future: First contribution. Proceedings, Club of Bologna, Oct 22–23. Available at https://www.clubofbologna.org/ew/ew_proceedings/2021_S1. 1_Gavioli_KNR_P.pdf. Kremmer M, Werner R, Masson BP, Mann JC. 2021. Method and machine for plant cultivation on a field. US Patent 11134608. Myers A. 1996. Method and apparatus for measuring grain mass flow rate in harvesters. US Patent 5,343,761. Oksanen T, Auernhammer H. 2021. ISOBUS—The Open Hard-Wired Network Standard for Tractor-Implement Communication, 1987-2020. Agricultural Equipment Technology Conf, Feb 8–10, Louisville KY. Pfeiffer D, Blank S. 2015. Real-time operator performance analysis in agricultural equipment: Understanding unused potential and ways to improve from day to day. Proceedings, Internatl Conf on Food and Agricultural Engineering, Nov 6–7, Hannover (pp. 359–64). Reid JF. 2011. The impact of mechanization on agriculture. The Bridge 41(3):22–29. Reid J, Moorehead S, Foessel A, Sanchez J. 2016. Autonomous driving in agriculture leading to autonomous worksite solutions. Proceedings, SAE Commercial Vehicle Engineering Congress, Oct 4–6, Rosemont IL (Paper no. 2016-01-8006). Schueller JK. 1992. A review and integrating analysis of spatially-variable control of crop production. Fertilizer Research 33:1–34. Searcy SW, Schueller JK, Bae YH, Borgelt SC, Stout BA. 1989. Mapping of spatially-variable yield during grain combining. Transactions of the ASABE 32(3):826–29. Sugumaran R. 2017. UAV docking system and method. US Patent 20160144982A1. Willrodt FL. 1924. Steering attachment for tractors. US Patent 1506706. [1] US Sugar has its own wi-fi network, spanning 270 square miles, at its Clewiston refinery in Florida (https://www.ussugar.com/technological-innovations/). About the Author:John Schueller is professor of mechanical and aerospace engineering, University of Florida. John Reid (NAE) is vice president, Enterprise Technologies, Brunswick Corporation.