Farming Drones: ~3x Yield Automated Precision Farming At Lower Costs

Summary:

The ratio between average and record yields for many crops is around 3x. This is roughly the yield increase that we can expect with fully automated precision farming that enables AI to almost fully optimize farming. This will globally disrupt farming, a ~$13 trillion/year industry. This requires regenerative farming practices, intercropping where multiple crops are grown at once so that no photon is wasted, and imaging and acting on the field on an almost daily basis. In this paradigm weeds never have an opportunity to grow and plants are tended and harvested individually, not collectively. Low value cereal crops can only afford the passage of a tractor two to three times a season, and robotic tractors, of whatever size, only marginally improve this. This is due to the inherent inefficiency and cost of off road tires. High intensity robotic tractor approaches are only viable on high value crops, which further limits economies of scale and corresponding cost reduction curves. 

Six axis farming drones bypass the need for tractors completely and can continuously monitor and tend a field, on a plant by plant basis, in a manner that is economically compatible with low value cereal crops. Six axis farming drones can automatically change farming tools and work independently of field size, shape, or slope. They do not compact the soil and they are of low enough cost that they can potentially be used for irrigation. Highly water efficient sub-surface injection directly to the root zone is feasible. Six axis farming drones enable us to fully automate farming and increase average yields by around 3x while also reducing costs and increasing resilience and food security. They can feed the world while enabling large scale rewilding of unused farmland to an extent sufficient to restore biodiversity and avert a mass extinction. Also removing enough atmospheric carbon, at negative effective cost, to stop global warming.

Concept:

Six axis farming drones ultimately perform all field farming operations, and should develop to this state quite rapidly, such that tractors and even people should never need to enter a field. They effectively convert a field into a large CNC machine with automatic tool changing and tools developed to perform every farming function. An analogy might be that they are a small, fast, low-cost, and autonomous flying tractor with implements for all field operations. Planting seeds and seedlings, monitoring plant conditions and growth, adding fertilizer, water, etc., and eventually harvesting the plant when it is ready. The field would be imaged at least daily and acting on that data would be continuous and almost immediate. Along the way poor performing plants might be removed and replaced, often by a different type of plant with multiple different plants being grown in the same area in a phased manner. This process helps to mitigate disease, pestilence, etc., and generally keeps the soil and growing conditions optimal and balanced so as to maximize overall yield and revenue, perhaps even without the need for spraying. Farming drones might also directly remove some pests and discourage birds, etc. Soil should never be exposed and photons never wasted with plants immediately replaced upon harvesting. Crop rotation and cover crops effectively being integrated in a continuous manner. Ideally non-direct revenue producing cover crops would never be needed. For example, cover crops and agricultural waste might be harvested for animal feed and/or biofuels with nutrients returned directly to the soil so as to maximize both revenue and soil regeneration benefits. Nutrients removed from the field via harvested crops need to be continually replenished, preferably by closing the nutrient loop. A six axis farming drone paradigm also enables a better integration with natural ecosystems, including rewilding, hedgerows, wildlife corridors, and so forth, that enable a more sustainable balance with multiple co-benefits.    

The six axis farming drone or drones might be automatically charged at the edge of the field where they would also have access to an array of tools for working the field. They would also have access to inputs, including seeds, fertilizers, etc., and access to a collection point for harvested crops. The drones might typically be quite small, likely under 55 pounds in many cases (an FAA drone class), although both smaller and larger drones might be used for different purposes. Multiple smaller drones can reduce environmental impact, including noise, and are more reminiscent of biological ecosystems. Because drones travel much faster than a tractor they can be much smaller while serving the same area, although plant by plant farming greatly reduces their effective speed - frequent stops. Farming drones might autonomously move between different fields as needed, often carrying tools, energy, supplies, and harvested crops with them. There is an optimization around the number of drone stations for a given area of farmland. A higher drone station density increases infrastructure costs but reduces average travel times between a drone station and the field. 

Solar energy, for example, via a solar hedge and a battery energy storage system, are integrated with the field so as to minimize energy costs. ~$0.03/kWh with battery storage should be possible, with ongoing cost reduction curves. An electricity grid connection is neither required or desired, as it would substantially increase energy costs and the need to distribute electricity around the farm, which can be expensive, delaying, and inflexible. Existing sunk cost grid connections might be used in some cases when lower cost energy was not available, for example, existing grid connections to irrigation pumps. A backup generator system, ideally powered via agricultural waste, might be used when necessary, for example, during a high energy use harvesting period. Although agricultural waste might also be a valuable biofuel feedstock and additional revenue stream. Six axis farming drones might also be used to carry batteries around a farm, enabling electricity to be economically and adaptively transported from where it might be generated, or where there is an existing grid connection, to where it might be needed at any point in time, avoiding the need for fixed power transmission infrastructure. Drone battery delivery can be an economically competitive replacement for power transmission systems for short distances of up to ten miles or so. 

Aircraft have been flown autonomously for decades, it is in many ways a much easier controls problem than autonomous self driving tractors and more comparable to industrial robots. High precision positioning systems such as real time kinematic GPS, ultra-wideband, ultrasonic, radar, lidar, and imaging systems might be used. Seeding and seedling planting might occur with this degree of accuracy (~10mm) resulting in a living plant map that can be used for automated farming activities including monitoring, fertilization, irrigation, and harvesting. Farming drones might scan a field at least once per day, continuously monitoring growth and enabling the very early detection and mitigation of any problems. In many cases scanning might occur in the process of performing other farming activities, although smaller scanning optimized drones might also be utilized within a drone fleet. Soil sensing would similarly be performed via drone mounted probes, as can under canopy imaging, direct soil and plant sampling, and so forth.   

The high level of physical data collection combined with the ability to act on that data presents a huge opportunity for the application of AI. This will enable substantial optimization of yield and revenue and the customization of performance to local conditions, including soils and climate, and the process of improving soils to much higher productivity levels. Another interesting loop that can be closed is between phenotyping and genetic optimization. Six axis farming drones enable the continuous phenotyping (measurement of critical plant metrics) of all crops and this can be combined with genetic information for far more comprehensive genetic optimization to local growing conditions, which can vary substantially over even short distances. Every farm effectively becomes an automated test site for plant breeding. This ultimately enables substantial further yield improvements and greater resilience.   

Six axis drone farming can directly integrate with greenhouses. Small six axis drones can operate within greenhouses, where they enable fast automation of operations, but this can also be integrated with outdoor farming. By operating a nearby greenhouse to germinate seeds and prepare seedlinglings for outplanting, yields can be significantly improved and effective growing seasons extended. This can substantially help with maximizing the utilization of farming area and available sunlight. It also improves the ability to phase different crops in and out, with seedlings able to immediately replace harvested plants. Greenhouse area costs substantially more than open fields, so there is a cost optimization that is dependent on crop value.  

Six axis farming drones require the development of a complete set of farming tools, the equivalent of farming implements that are towed by tractors, but at high speed drone scale and weight with the capacity for individual plant level operation. For example, tools for seeding, fertilizing, seedling outplanting, monitoring, irrigating, soil sensing, harvesting, stalk removal and bailing, and so forth, and these tools need to be customized and developed for a wide range of plants. The highly paralyzed large farming implement being replaced by many small individually controllable tools and drones. The precision flying capability of six axis drones enables the drones to directly place and pick up farming tools from a tool holder at the side of the field, and these tools might be traded with neighboring fields. Similarly battery charging and/or battery swapping is accomplished directly via the drones, with drones accurately landing on electrical connectors for charging or batteries for battery swapping. Charging stations being connected to and in some cases directly integrated with colocated solar power systems. Farming is very cost sensitive and so is six axis drone farming.  

Irrigation is a substantial farming cost, both in water use and infrastructure needed to apply it efficiently. For example, large center pivot irrigators or sub surface drip irrigation systems might commonly be used. Six axis farming drones can be sufficiently low-cost that they can be used directly for transporting water very short distances without the fixed infrastructure of pipes. This is further aided by the use of a water tower. The irrigation drone can pick up water at height and effectively glide down to the field using minimal additional energy. With a map of the position of every plant and the use of a probe that can be inserted into the soil for precise water injection to the root zone, six axis farming drones can be very water efficient. In a water limited field, ideally all water that is evaporated from the field should be evaporated via plants, passing up through the roots to get there. There should never be any exposed wet dirt. Another advantage of high yield farming is that water use often does not increase in proportion to yield. Higher yields can enable lower water use per unit of yield. High precision individual plant level farming also enables greater optimization for reduced water use. Six axis farming drones enable a highly precise, water efficient, and adaptable approach to irrigation without the need for substantial and labor intensive infrastructure.    

Economics:

One of the fundamental cost driver advantages of six axis farming drones over robotic tractors is the efficiency of transport. Six axis drones can achieve transport costs comparable to trains, which is substantially lower than on road trucks, which are, in turn, substantially lower than off road tractors. Transport costs are typically around a third each for capital, operations and maintenance, and energy or fuel costs. Six axis farming drones can have very low capital costs because they can be long-lived with very high speed operation and utilization. A tractor can last up to 10,000 hours, which if operating at an average speed of 15mph, translates to around 150,000 miles. In contrast, a passenger jet can last more than 100,000 hours or 50,000,000 miles. While a farming drone may not travel as fast as a passenger jet and might become obsolete much sooner, utilization based longevity can still effectively be more than 10x that of a tractor. The same general relationship applies to operations and maintenance costs, noting that six axis drones might have as few as six long lived moving parts. Drone based farming tools are unlikely to be substantially different to farm implements in terms of longevity and operations and maintenance, although automatic tool changing enables quick replacement and tools can easily be automatically transported to and from a maintenance location. Drone farming tools are also more subject to mass manufacturing and associated cost reduction curves. Co-located solar can be around a tenth the cost of diesel, so there is also a ~10x saving on energy costs. Electrified tractor robots using colocated low-cost solar energy would also benefit from this, although they still suffer from high capital and operations and maintenance costs, not to mention other technical limitations like terrain dependence, soil compaction, plant height compatibility, water transport, and so forth.   

Assuming yields commensurate with six axis farming drones minimum annual revenues for low value cereal crops might be on the order of $3,000/ha. While there will likely be a diversity of drones collectively operating over larger areas, a 55lb class six axis farming drone might on average farm around 10 hectares by itself. This is mostly set by peak harvesting requirements. This drone might cost a few thousand dollars and last around five years on average. Noting that six axis drones might typically be quite robust with high payload fractions, strong wind tolerance, and minimal moving parts. The solar system to power it might be closer to $10,000 and the farming tool set and irrigation water supply would also likely cost more than the drone, but farming already incurs most of these costs. A few more drones per hectare would be needed to cover peak irrigation loads, especially in low rainfall areas, but again this does not need to be a dominant cost. Irrigation drones which might partially glide down from a water tower could be optimized for the task and have higher payload fractions. All up the payback periods for six axis farming drone based farming, with its ~3x yield/revenue increase, could be as little as one to two years, although it may take a few years to build up soil quality and achieve high yields in some cases. Payback periods could be even less if some proportion of higher value crops could be included. 

Another interesting consequence of ~3x higher average yields at lower cost is that it becomes less expensive to increase production by increasing yield on an existing farm than by buying the farm next door. Farmers can expand by ~3x without purchasing more land. This leads to substantial economies of scale, as an existing farm can generate ~3x the revenue without the additional logistical constraints of tending more land, even as those logistical constraints become less constraining enabling yet larger farms. Farming can become much more profitable. Around 26% of workers globally are employed in the agricultural sector. More food will be produced and it will cost less, but care must be taken to avoid a substantial dislocation of workers. A large fraction of traditional farms will be rewilded, and necessarily so if we are to avoid a mass extinction and global warming. One method will be to integrate biofuel production and natural habitat management into this farming transition process. This will substantially increase total revenues beyond current agricultural sector revenues, displacing a substantial fraction of the existing fossil fuel industry and generally improving the environment with corresponding co-benefits. This should provide more and better employment opportunities in excess of the current agricultural industry.

Six axis drone farming will likely first be used on high value crops, as that is where the profit margin is greatest. Care has to be taken to not become trapped in high value niche applications, as farming scale and cost reductions will be claimed by the much larger scale lower value crops, which might then include higher value crops in their process at little additional cost. Six axis farming drones have the potential to greatly reduce the cost of currently higher value crops. Among other things they can automate a lot of processes that currently depend on human labor, including picking. Having high speed and being relatively reach independent they are particularly well suited to pruning and harvesting trees. They also do not need ground vehicle access pathways, significantly increasing the utilization rate of available land. Many high value crops will require the development of specialist farming tools, which can slow the deployment of these systems. For example, tools for carefully selecting and picking berries and fruit.      

Impact:

With the ability to fully automate farming and increase average yields by around 3x while significantly reducing the cost of food production, six axis drone farming will be truly disruptive to the agricultural sector. Especially as they can scale incredibly rapidly with a minimum of additional infrastructure. Six axis farming drones are conducive to automotive type mass manufacturing with the ability to quickly ramp up production into the millions of units per year. Accelerated by the AI revolution this might enable a substantial farming transition in as little as 10 years - by then a ~$30 trillion per year opportunity. Quickly surpassing autonomous operation and electrification of traditional farming methods. Scaling rate is now almost everything and the technology that can scale the fastest generally wins. It is also a leap frog technology that enables high performance farming largely independent of prosperity level. Like cellphones and electric vehicles, farming drones can operate effectively in low and high income countries alike, and have a huge impact on the prosperity and food security of both lower and higher income countries.

Beyond enabling higher value, lower cost food for all, this technology also has dramatic environmental benefits. We farm around half of the world’s productive land and with much higher yields we can return a large fraction of this to nature. We find that restoring 15% of converted lands in priority areas could avoid 60% of expected extinctions while sequestering 299 gigatonnes of CO2—30% of the total CO2 increase in the atmosphere, or 14% of total emissions, since the Industrial Revolution. This restoration can avert a mass extinction, noting that it takes millions of years to recover biodiversity after a mass extinction and so this is of even greater importance than mitigating climate change. However in doing so it will also remove enough carbon dioxide to stop global warming. There is a tipping point associated with six axis drone farming, once this tipping point is reached, and the cost of reaching this tipping point is likely less than a billion dollars, the transition becomes economically self-driving. That is, beyond this tipping point, the cost of restoring biodiversity and removing carbon dioxide at magnitudes great enough to stop global warming, becomes negative - six axis farming drones pay for themselves. There are few if any technologies more critical to saving the planet than six axis drone farming. Ocean farming, low-cost desalination for desert farming, and highly efficient biofuel production are highly complementary, but high intensity farming via six axis drone farming is likely the most critical of all of these. Averting a mass extinction while feeding the world is the first priority.