How autonomous vehicles will improve sustainability and productivity in agriculture

This article originally appeared in the June'23 magazine issue of Electronic Specifier Design – see ES's Magazine Archives for more featured publications.

While drones and agricultural robots represent new systems being developed and deployed from scratch, tractors are different. There’s already a large installed base of tractors, and they tend to have long operational lives. As a result, in addition to developing fully automated new designs, existing tractors will be retrofitted with electric drives and upgraded with digital systems for specific purposes, so-called ‘digital tractor implements’.

This article, by Rolf Horn, Applications Engineers, DigiKey, looks at the development of digital tractor implements and emerging electric tractors (e-tractors). It examines some of the technologies needed to realise the development of autonomous agricultural vehicles and how Digi-Key’s product offering can help designers speed their development processes.

Farm implements get on the ISObus

Like Industry 4.0, agriculture is moving toward using intelligent and interconnected machines. That’s where International Standards Organisation (ISO) 11783, the tractors and machinery for agriculture and forestry serial data network bus, comes in. In the agriculture industry, it’s simply referred to as the ISObus. It’s based on the Society of Automotive (SAE) J1939 protocol, which includes the control area network (CAN) bus and has been optimised for agricultural applications.

The ISObus includes standardised connectors, communication protocols, and operational guidelines and enables the development of interconnected sensor and control systems from different makers (Figure 1). ISObus also supports the electrification of tractor implements, including electrically driven mechanical power take offs (PTOs) and high voltage connectors rated for up to 700 volts (V) and 100 kilowatts (kW) to power electrically driven implements.

Figure 1. ISObus can enable the integration of sensors and implements from diverse makers into a plug-and-play system. (Image source: Photo by Kinwun via Getty Images)

The ISObus is evolving to develop a tractor implement management (TIM) system. As envisioned, the advanced version of the ISObus will enable implements to provide feedback to the tractor, supporting the optimisation of the combined tractor/implement system. It will also allow higher levels of sensor integration on implements supporting precision farming. The tractor will provide location awareness, and the combined system will continuously gather data about the soil and crop conditions. With more detailed insights, yields and sustainability can both be increased.

E-tractors, retrofits, and autonomous tractors

E-tractors are beginning to appear. In addition to reducing emissions (a quarter of the world’s greenhouse gas emissions come from agriculture and agriculture-related activities, and one tractor is equal in emissions to 14 cars), e-tractors can significantly reduce fuel costs. E-tractors are currently limited to smaller models as large, high-power e-tractors require battery packs larger than the size of the conventional tractor they would replace. Large e-tractors also weigh more, resulting in increased undesirable soil compaction. Finally, charging times for large battery packs are too long to be practical in an agricultural operation. Smaller e-tractors with motors from 25 to 70 horsepower (HP), about 18.6 to 52kW, and small battery packs are already being tested. Tractor electrification is about more than the drive train. It’s also about replacing hydraulics for powering and controlling tractor implements (Figure 2).

Figure 2. Small e-tractors with motors from 25 to 70 HP are being tested and ready for deployment. (Image source: Photo by brizmaker via Getty Images)

For larger tractors, hybrid retrofit kits are available. For example, one company offers a kit with a 250kW generator that can be attached to the tractor’s existing internal combustion engine in place of the hydraulic pump. The kit also includes four electric motors to replace the hydraulic drive system and an electric transmission to power existing implements. By replacing the hydraulic systems, the retrofit kit reduces fuel and maintenance costs and increases the availability and reliability of the hybrid e-tractor.

Like the rollout of autonomous automobiles and trucks, the deployment of autonomous tractors faces an indeterminant future. For example, current regulations in California require that “all self-propelled equipment shall, when under its own power and in motion, have an operator stationed at the vehicular controls”. Full autonomy will have to wait.

Flying over the fields

Drones are currently used for a wide range of tasks in agriculture. Examples include:

• Imaging plant health. Drones have largely replaced satellite imagery to monitor crop health. Equipped with Normalised Difference Vegetation Index (NDVI) imaging equipment, drones provide detailed colour images that can be used to monitor plant health.

• Monitoring field conditions. Drones also monitor the soil and drainage conditions over entire fields. That can enable more efficient and sustainable watering programs.

• Planting. Automated drone seed planters are common in forestry industries, and their use is being expanded to general agriculture. Drones can rapidly plant trees or seeds and reach inaccessible areas more efficiently.

• Spray applications. The use of drones to apply spray treatments of fertilisers and pesticides is an emerging application whose use varies by region (Figure 3).

Figure 3. Large drones have been developed that can be used to apply spray treatments of fertilisers and pesticides. (Image: Photo by baranozdemir via Getty Images)

Precision produces more with less

Even before autonomous tractors are realised, drones and the electrification of tractors and tractor implements are expected to support precision agriculture and increase sustainability.

According to a study by the Association of [agricultural] Equipment Manufacturers (AEM), the use of precision agriculture can lead to a 4% increase in crop production, 7% increase in fertiliser placement efficiency, 9% reduction in herbicide and pesticide use, and 6% reduction in fossil fuel use. In addition, water use can be reduced by 4% with precision irrigation.

Those numbers are based on current technology. With the addition of connected systems and artificial intelligence (AI), those improvements are expected to be multiplied. Adding machine learning (ML) for equipment maintenance provides further savings and improvements in sustainability. According to the AEM, autonomous farm equipment is expected to result in an incremental 24% improvement when both input savings and yield improvements are considered.

AI and ML will also be critical for developing precision machines optimised for specific tasks. Dedicated task machines can be even smaller than general-purpose tractors. For example, small task machines are being developed for picking crops where machine vision, a delicate touch, and precise dexterity are required (Figure 4).

Figure 4. Example of harvesting autonomous robots that combine machine vision with AI and ML. (Image source: Photo by onurdongel via Getty Images)

Weed control is another area where task-specific AI and ML machines are expected to contribute significantly. Weed control is difficult, labour intensive, and, if not efficiently implemented, contributes to the use of more water and depletion of soil nutrients. Crop rotation is a partial solution but cannot eliminate the need for herbicides or manual weed control. Weed management robots that combine machine vision with AI and ML are being tested. These small machines also minimise soil compaction.

Farm OS and fleets of autonomous equipment

The agriculture industry is looking toward a future where fully autonomous farms will be controlled by a sophisticated operating system (OS) capable of managing mixed fleets, including both autonomous and standard farm equipment, plus land-based machines and drones, to maximise productivity and sustainability (Figure 5). Those fleets will be operated in coordination to help control capital expenses, minimise labour needs, and provide the big data necessary to enable autonomous execution and precision agriculture. In addition, the farm OS of the future will be standardised and optimised to support a diverse range of equipment from numerous suppliers. Adopting the ISObus is only the first step toward an open-source and standardised approach to farm automation.

Figure 5. Swarms of coordinated ground and flying autonomous agricultural machines will lead to higher levels of sustainability. (Image source: Illustration by Scharfsinn86 via Getty Images)

Additional benefits expected from the proposed farm OS are reduced CO2 emissions, lower fuel consumption, and optimisation of battery charging and management. Big data analytics will also play an important role in the future of agriculture. Large amounts of real-time data directly from the field will be used to continuously train the AI and ML algorithms required for decision-making, control, and operational planning to optimise precision agriculture.

Summary

It’s still in the early days for the development of autonomous farm vehicles and sustainable precision agriculture. The industry has started down the path with ISObus. The next generation of ISObus will support increased interoperability and help lead to more complex and interconnected fleets of farm equipment. The goal is the development of a farm OS that can take those fleets of farm equipment, combine them with massive real-time sensor data using AI and ML algorithms and deploy them as formations of coordinated ground and flying machines producing high levels of sustainability and productivity.