When developing products for smart agriculture applications, supply chain forward thinking pays dividends, as does partnering with a silicon vendor with expertise in other sectors in which AI/ML and smart sensors are being employed.
IoT Revolutionizes Smart Farming Practices
The twin goals of smart agriculture (a.k.a. smart farming) are yield improvement and higher profitability. These goals are met by increasing both the quality and quantity of crops (as well as livestock and dairy products) while optimizing the use of inputs, such as seeds and pesticides. However, such optimization requires the increased use of the internet of things (IoT), with an IoT Business News article from late 2023 observing “…with IoT technology, smart agriculture is changing the way we cultivate and manage crops and livestock. IoT involves the connection of sensors, devices and machinery to the internet, enabling data collection, analysis and real-time decision-making.”
The IoT effectively provides crucial connectively within a layered smart agriculture ecosystem.
Figure 1 – Smart agriculture futuristic industry
The smart agriculture physical layer is predominantly IoT-enabled devices, monitoring a variety of parameters such as temperature, humidity and soil and crop health. Some sensors will be in or on the ground, others will be airborne (fitted to drones) and others will be within RFID tags attached to livestock.
Data is typically transmitted real-time to the edge-layer for immediate analysis. It is then passed on to the cloud layer for deeper analysis and for decisions to be made against weather forecasts and the real-time costs of inputs and prices being paid for crops, for example.
Artificial intelligence (AI, or machinery mimicking human intelligence) is used extensively at the cloud layer whereas machine learning (ML, which is more about pattern recognition and inference from data), is used at the edge layer and, increasingly, at the physical layer in the form of “smart sensors” that use edge-processing.
The decisions made (by humans as well as AI) on the collected data can be used to provide instructions to automated farm machinery to, for example, manage/maintain soil quality, plant seeds where they will produce the best results and apply water and pesticides only where necessary.
Moreover, real-time data collected and analyzed by the automated machinery can be used to make immediate decisions. For example, mobile machinery with ML-based camera systems can recognize the difference between plants and weeds and apply appropriate treatment as it moves through fields. Indeed, this real-time and localized analysis and action is so-called precision agriculture (PA) at its best, with the US Government Accountability Office (GAO) saying “precision agriculture technologies can improve resource management through the precise application of inputs, such as water, fertilizer and feed, leading to more efficient agricultural production.”
Essentially, we have made possible an extremely high-tech ecosystem, but in my opening gambit I said “profitability” was a goal.
Product Development
Understandably, the more data that can be captured at source and the more actions that can be automated the better. There is therefore a great need for IoT-based systems in the field and for automated machinery that also uses the IoT. Indeed, the IoT in Agriculture Market is projected to “…grow from USD 18.43 Billion in 2024 to USD 71.75 billion by 2032, exhibiting a compound annual growth rate (CAGR) of 18.52% during the forecast period (2024–2032)” (Shubham Munde).
From a product development perspective there are many technical challenges to address. Most sensing systems will be in remote locations, so they will need to be battery-powered (supported by solar cells), and mesh network technologies will most likely be needed to get data from the physical layer to the edge and layer. Systems will need to be rugged to operate in harsh environments and, ideally, not require maintenance once put into service. As for systems fitted to machinery or vehicles, they will need to contend with high vibration levels, mechanical shocks and possibly dust.
These are not insurmountable challenges for an industry that is capable of getting equipment to work in space and on other planets. However, smart agriculture needs cost-effective solutions and for projects to be de-risked as much as possible. Accordingly, smart business and engineering practices must be employed throughout the product development workflow and once the product enters service (see Figure 2).
Figure 2 – The IoT product development flow
Regarding the flow, it is worth noting that bringing in partners occurs immediately after the product’s requirements and functionality are set, though in reality there may be a need to take a step back and revisit the first stage. Indeed, there’s an argument for including partners as the product’s requirements and functionality are still being defined.
For instance, third parties with existing experience in smart sensing, AI, ML, low power electronics and RF can provide invaluable help when it comes to device selection. In this respect, Microchip is an ideal partner because we have devices—including microcontrollers (MCUs), microprocessors (MPUs), FPGAs, Ethernet solutions and smart sensors—that are not only used in smart farming but also proven in automotive and industrial applications.
Skills Shortages and The Supply Chain
Partnering with Microchip can also help those developing products for smart agriculture applications overcome two major challenges currently facing the engineering sector.
Firstly, there are global engineering skills shortages. A recent Boston Consulting Group survey revealed that 92% of leaders of global companies rank attracting and retaining talent among their top three priorities, and a BCG article (developed in partnership with SAE International) that cited the survey noted “engineering talent in particular is in short supply, with a growing gap between supply and demand that puts the progress of important industries at risk and threatens to have a tremendous negative impact on the economy—equal to nearly 40% of the projected GDP impact of all talent gaps expected in the US through 2030.”
Not surprisingly, OEMs are employing engineers straight out of university. While the graduates will have core engineering qualifications, they will have no practical experience. Also, while there is a drive to see more AI/ML-enabled automation within agriculture equipment, the fact of the matter is that that won’t happen overnight. Those fresh from university will have little, if any, understanding of the agricultural applications for which products, systems and equipment are being designed.
As mentioned, smart agriculture products will be operating in a harsh environment. Even so, they will be expected to be reliable and provide several years of service. But failures will occur, and either replacements will need to be sought or repairs made. Either way, obsolescence must be managed; a recent Converge Future of Obsolescence Management (FOM) LinkedIn poll (see Figure 3) rated four factors that contribute to obsolescence in the electronics industry.
Figure 3 – A LinkedIn survey weights four reasons for component obsolescence.
Understandably, the pace of change in the electronics industry accounts for the largest two contributors. As for the supply chain challenges, designers need to feel confident that parts will be available when they are needed—for prototyping and when going into production. In this respect, our supply chain specialists, who are available to attend key account meetings, have the expertise to advise on mitigating the risks associated with rapid technological advancements and short product lifecycles.
Empathy
Though the market for IoT in smart farming is huge, and those who have embraced practices such as PA are seeing a return on their investment, it must be appreciated that farmers seldom have the funds to make a substantial upfront investment; they typically operate on very tight margins. That said, farmers can take advantage of funding opportunities and government-backed financial incentives.
Also, those supplying products should present a migration path that allows for expansion of their smart agriculture ecosystem over time. Interoperability and backwards compatibility should be key. Again, Microchip—renowned as a key player in the world of embedded systems—can recommend which hardware products will be most suitable. Moreover, we have several design partners specializing in sustainable applications and reference designs.
It is also important that all smart agriculture products be secure as, with the IoT expanding, the “attack surface” is growing for hackers. Data paths exist between devices in the field and the cloud. Also, for many of the products to have longevity they must be able to accept software upgrades. In this respect, we have an extensive portfolio of security products. These include MCUs and MPUs with integrated security, platform root of trust controllers, authentication ICs (with secure key storage and hardware cryptographic accelerators) and secure FPGAs.
Figure 4: Security is key
However, it is important for ease-of-use to be built into all smart agriculture products. Farmers are busy enough without having to become IT experts. There is, unfortunately, a risk of lack of standardization within the smart farming sector as companies race to introduce products. Again, we can help—by recommending solutions that have the highest levels of interoperability, but which are no less secure because of it.
In addition to the aforementioned (meeting low) power draw challenges, we can also advise on size and weight, a factor that is particularly crucial for drones and the equipment they carry. Meeting size, weight and power (SWaP) requirements is standard for all embedded systems.
Summary
Farmers need to switch to smart agriculture practices if they are to become more profitable, and as an added benefit it is a more sustainable form of farming as fewer resources and inputs are used. Switching means embracing a raft of new technology (predominantly IoT) that will likely feature ML. Accordingly, smart agriculture IoT is big business, and products are entering the market all the time.
Microchip makes an ideal partner because although smart agriculture is a relatively global new market, we are already selling products into it; products that were designed with low power, high performance, security and AI/ML in mind, and for which development kits exist in many cases. Moreover, our supply chain specialists, working on key accounts, help developers work around any challenges that arise during the development and out—often literally in the field.
Smart agriculture requires smart business practices, and one of the smartest things to do is work with smart partners.
Brad Poole, Nov 7, 2024
Tags/Keywords: Industrial and IoT, Sustainability
Comments