CROSSING THE CHASM

For decades we’ve been hearing that we will eventually cross a chasm and computers, software and data analytics will finally form-fit themselves to agriculture and farming. And yet, more than twenty years after the introduction of GPS technologies on farm implements, adoption of smart farming and precision agriculture systems remains uneven and slow.

With the advent of the industrial age, when society changed from an agrarian base to an industrial base, the way humans worked became more mechanized. Industrial nations built factories and furnished farms with new mechanical tools and systems that were very different from those used in the agrarian era: tractors, steam powered engines, the railroad, telegraph, and much more formed the basis of modern society and business.

Observers have said for more than twenty years that the impacts the information revolution will have on agriculture and food production will be far greater than the impact science had during the transition from an agricultural society to an industrial society. If this is true, then why has the agriculture and farming sector moved so slowly to adopt new digital and information technology?

Farming is under extraordinary pressure from many diverse forces including climate change, the need to produce food sustainably, falling commodity prices and the rising debt of farmers. This phase of technology evolution – Smart Systems, Services and the Internet of Things (IoT) — is supposed to be setting the stage for a multi-year wave of innovation and technology adoption in the farming arena. But is it?

SMART WHAT?

In its simplest form, smart agricultural and farming systems is a concept in which inputs—from plants, soil, animals, sensors, machines, people, video streams, maps and more—is digitized and placed onto networks. These inputs are integrated into systems that connect farmers, machines, processes, and knowledge to enable better decision making for agricultural systems and food production.

Whatever we chose to call it — “Smart Farming” or “Precision Agriculture” or some new spin on “Industrie 4.0”— we are referring to the practice of utilizing digital and software technologies to manage the spatial and temporal variability of growing crops to improve the efficiency and sustainability of food production. For precision agriculture to reach its promise, certain challenges will need to be met, including:

Standardization Across Equipment Platforms

Equipment OEMs are at the forefront of smart farming systems helping to drive the adoption of many new innovations. However, standardizing equipment standards and interfaces across diverse OEMs will be a minimum requirement if farmers are to realize the value of these new innovations.

Connectivity Challenges

In developed countries, the majority of food production is shifting to mid to large-scale farming operations. In the US in particular, less than 8% of farms account for over 60% of total production. These larger farms require extensive network infrastructure to provide connectivity across operations.

For both small and large-scale farms, connectivity challenges have continued to be persistent. Rural internet access leaves much to be desired. This is particularly true in developing economies, but it’s also true in developed regions like North America. There are many places on Earth where dropping a hard drive in the mail is more efficient than uploading data on a weak 3G signal. The lack of dependable connectivity (4G or greater) will continue to hamper adoption of new digital systems.

Data Management at Scale

Data and analytics are the core value creation mechanisms within smart farming systems. Today. the data being collected from farming implements is exploding which, in turn, is setting the stage for analytics, AI and machine learning solutions.

Farmers increasingly will need to understand which data points they value and in what time frame, daily, monthly or seasonally. Trying to manage the rapidly expanding volume of data as it accumulates is a monumental task, never mind the challenge of processing the data in real time.  Even a small farm has hundreds of thousands of different data points that potentially can be collected. Today, it’s still far easier for farmers to enable data collection from modern farm machinery than it will be to aggregate, analyze and gain insights from the data.

Manufacturers like John Deere and Case collect and store data in proprietary formats that must be converted into common formats where the data can be shared and analyzed. Evolving standards like ESRI Shapefiles, and more recently GeoJSON and GeoPackages, provide open standards for how this data can be shared and analyzed, but in the conservative culture of most farm equipment OEMs, competitive advantage is usually perceived, to one degree or another, to lie in ownership and control. It goes without saying that such a culture does not blend well with the notions of openness, transparency and trust.

Farmers expect evolving software tools to be functional, ubiquitous, and easy-to-use. Within this construct, however, the first two expectations run counter to the third. In order to achieve all three, a new approach is required. Achieving completely fluid information and fully interoperating devices, data, people and systems—requires an equally simple, flexible, and universal abstraction that will make information itself truly portable in both physical and information space, and among any conceivable farm implement or technology.

AI and Machine Learning Inform Autonomous Adaptive Solutions

High performance networks, edge computing, data analytics and machine learning are setting the stage for complex adaptive or autonomous systems in agriculture. Similar to self-driving cars, the ability to process large scale data to model processes and derive algorithms that successfully shortcut calculating information will revolutionize agriculture. For example, the ability to combine image processing (including telemetry and LIDAR) with commercial drone technology where data is far more efficiently collected from the air than the ground (including elevation, slope, weed prevalence, pest prevalence, etc.), will drastically improve the efficiency of optimizing crops. A farmer will be able to canvas a farm quickly and at a substantially lower fuel cost than driving a tractor across the farm to do the same tasks.

Scalability of Solutions

Agriculture can exist at almost any scale, from one-man operations to powerful corporate farming enterprises, but many of the challenges farmers face no matter what the scale will be similar. For precision agriculture to succeed, application solutions will need to be scalable. Farmers will need to utilize the same digital tools for both large and small farms.

The Future of Farming Is Heavily Dependent on Ecosystem Development

Smart farming and precision agriculture is a crowded field. The opportuni­ty is informed by diverse players with differing business models including ag­ricultural equipment manufacturers, seed suppliers, chemical companies, software providers, wireless carriers, IT systems and cloud computing services, dealers/distributors, services providers and many diverse ancillary services (financing, investment capital, insurance, etc.). Conse­quently, farmers and agriculture industry participants are faced with rapidly changing technology and a supplier land­scape that is globally dispersed with poorly developed ecosystems.

This ecosystem fragmentation reflects the state of data management in agriculture, where inputs to a central farm management system (including crop, environment and machine data) remain isolated in functionality. Sensors, devices, drones, combines, storage silos, robotics and more will produce billions of data points about how crops are planted, grown, protected and harvested, increasing visibility and precision in all agricultural enterprises. All this data, however, requires significant integration and data aggregation and management expertise for it to be utilized effectively—a truly open data ecosystem is needed to access the full scope of value created by these smart, connected systems.

Once data is accessible, artificial intelligence and machine learning algorithms can help farmers track crop health and trends to make critical decisions. AI/ML can help farmers manage precise variability rather than making decisions based on algorithms. How deep should I plant this genetic variation of corn? How low should I fly my crop duster today given wind speed and direction? And on and on.

OEMs like Deere and AGCO, and feed, seed and chemical suppliers like Monsanto are all trying to be the center of this new agricultural data ecosystem, but are focusing on their own offerings, largely failing to catalyze wider adoption. Data platforms that integrate information across devices, systems and farms will win the day, especially if farmers retain ownership of their data.

The benefits of large-scale collaboration, ecosystems and developer communities in the smart farming and agriculture arena are just beginning to be recognized. What would a precision ag data ecosystem look like and what potential could such ecosystem and collaboration inform? How should innovative players, large or small, engage in new collaborative data relationships to drive market development?

Data will never come from a single unified source. What technology players looking to leverage data collaboration—or benefit from connecting diverse “smart systems” need to understand is that we have entered a phase in the marketplace where data with real practical value can originate from many sources across the farming ecosystem. It simply needs to be better organized, facilitated and orchestrated.

The collection of dull and dreary “solo” solutions—like equipment automation, fleet tracking applications and GPS-enabled location services—that comprise a significant percentage of the precision agriculture world today are really simple applications. These simple applications don’t really need to be “open for data sharing.”

So how will smart farming data ecosystems develop?  What’s really required to drive the aggregation, sharing, and data analytics innovations that have the potential to inform new collaborative business models?  This involves three interrelated elements:

1. A vision for how data collaboration networks will drive “catalytic” innovation to help focus participants on the value data can create;
2. An architecture that organizes open data, provides easy to use tools and support for analytics and reduces the investment and effort required to participate in data ecosystems;
3. Relationship enablers and economic incentives which persuade participants that the ecosystem developer is serious and can really scale data sharing and new services delivery systems.

Many farmers today still use many of the same production methods their parents and grandparents used, are independent by nature, and are often reluctant to look beyond long held traditions for new solutions to their challenges. These characteristics have been further aggravated by a growing distrust with large corporate agricultural players which have exploited small farmers in areas such as seed patents, profits, and ownership rights. Traditional culture and trust issues are the perfect ingredients to create resistance to new innovations. ◆

Stream our latest Future Perfect Tech podcast