Andy Pye gazes into the field of dreams to see how autonomous agricultural equipment is set to make light work of the harvesting season.
As stories emerge of crops rotting in the fields for lack of immigrant workers to harvest them, what place does robotics have for stepping in and saving us? According to Dr Khasha Ghaffarzadeh, Research Director of IDTechEx. agricultural robotics can upend several commonly-held notions. Amongst them is the idea that big is better. In practice, this has translated into ever larger and more powerful agricultural machinery. This makes sense because a big machine amplifies the capabilities of the skilled driver, dramatically boosting its productivity. This notion may however be about to experience a fundamental change. The reason is that the driver can now finally be taken totally out of the equation.
The IDTechEx Research report “Agricultural Robots and Drones 2016-2026: Technologies, Markets, Players” develops a detailed roadmap of how robotic technology will enter into different aspects of agriculture, how it will change the way farming is done, how it becomes the future of the agrochemicals business and how it will modify the way we design agricultural machinery.
The autonomous tractor
Agricultural vehicles have been at the forefront of developing and adopting autonomous navigation technology. Indeed, more than 320,000 tractors equipped with auto-steer or tractor guidance were sold in 2016 alone, expected to rise to 660,000 in 2026. These tractors use RTK GPS technology to autonomously follow pre-planned paths with centimetre-level accuracy. This makes agriculture the largest adopter of autonomous navigation.
Leading tractor companies worldwide have already demonstrated master-slave or ‘follow-me’ unmanned autonomous tractors or load carts. In these arrangements, a manned operator supervises the movement of the lead tractor with others following suit.
This technological evolution will further the notion that big is better, because it enables further amplification of the productivity of the skilled driver via multiple slave or follower vehicles. This arrangement will find increasing use in large-scale crop field farming. Fully and unmanned autonomous tractors will be the next evolutionary step. Multiple semi-commercial prototypes have already been demonstrated by some agricultural machinery companies.
The technical challenges are largely resolved. Here, the tractor becomes equipped with a variety of overlapping sensors such as LIDAR, RADAR and sonar to provide autonomous navigation in the absence of a GPS signal together with collision avoidance.
Technology costs are currently high but the largest hurdles are to be found in the lagging regulatory framework and the farmers’ desire to stay in charge. These will all inevitably change, particularly as the farming population further ages across the globe.
Mobile agricultural robots
Meanwhile, the advent of mobile agricultural robots will create the notion that small, light and slow is good. In this vision, a few heavy, fast, large and manned tractors become replaced by, or complemented with, many light, slow, small and unmanned robots.
Here, the lightness means no soil compaction, thus increasing the useful land in each farm by as much as 3%, and slowness means more attention paid to each plant – therefore better data and more precise plant-specific action. Small size also means potentially low cost.
However, because large and fast machines are more productive, these new classes of agricultural robots will need to be lower in cost by as much as 24 times to make economic sense and to enable mainstream adoption in the medium-term.
This would be a radical shift in the way we envisage agricultural vehicles. The emerging alternative vision is still in its infancy, but the direction of development is clear. There are already hundreds of mobile agricultural robots in existence. With the exception of 50 or so small-sized ones, most are however still in the research or semi-commercial trial stage.
The costs are currently high too, mostly because such mobile robots require multiple sensors to provide safety and autonomy even in the absence of GPS signals. The early evidence is that farmers do not yet trust them and indeed are not willing to pay extra. This means that some models are being stripped down to the bare minimum required with seemingly simple features such as even end-of-row navigation in orchards removed.
The machines are still not completely reliable despite the technology being ready and accessible at the discreet component and software level. All this is predicted to change in the coming decade. Thus far, only a few field trials have taken place and here the experimental clock is inevitably limited by harvest seasons, further slowing down the adoption process.
Asparagus is one of the most expensive vegetables in Europe because harvesters have to painstakingly pierce each stalk individually. A robot could change this, and engineers at the Bremen Centre for Mechatronics (BCM) are developing one. It works with harvesting tools, which run on precision rails from HepcoMotion, a specialist in linear guidance systems.
Asparagus has been steadily growing in popularity among British consumers, yet for the farmers it means spring time stress. This is because they have to harvest enormous quantities in such a short time – in 2015 according to the Federal Bureau for Statistics it was 112,100 tons. Work in the fields is arduous and workers have to cut through each individual stalk.
The green asparagus harvesting robotic system project (Garotics) is developing a harvesting robot for green asparagus. The Bremen Centre for Mechatronics (BCM), the packaging machine manufacturer Strauss from Buxtehude and the British agricultural company C Wright & Son are involved.
The basis of the harvester robot is a chassis with four wheels and a front-wheel drive. In the centre between the front wheels there is a camera system installed, which films the green asparagus stems as it goes past. Unlike white asparagus, green asparagus grows above the ground. Photographic processing software then identifies stalks that are ripe for harvesting. “One of the challenges was to implement image processing which can differentiate the different stages of growth,“ says Strauss design engineer Sebastian Allers.
Linear guide rails allow exact positioning of the harvesting tool. Software then guides the coordinates of the ripe samples further to the tool head, which is mounted under a hardened and precision-cut linear guide rail made of stainless steel. The tool head can travel across the full vehicle width from side to side on a timing-belt-driven carriage.
Drones for agriculture
French manufacturer Drone Volt is a company expanding the use of drones in agriculture. The DV WING flying wing drone is dedicated to precision agriculture and construction work. The French drone market for precision agriculture and mapping is estimated at around EUR50 million out to 2025, according to Global management consultancy, Oliver Wyman. On a global scale, the market potential for commercial drones servicing infrastructure, agriculture and mining operations is estimated at $81.0 billion between now and 2020 (source: PwC Survey – May 2016).
The DV Wing is a fixed-wing unit equipped with an 18.2 MP sensor and uses algorithms enabling it to obtain aerial imagery and accurate data for missions such as photogrammetry, map analysis for farming areas and forests, and measurements for road construction. It can also be used by quarry and mining operators to measure volumes.
The data it collects can be used by farmers to establish accurate diagnostics for the treatment of crops and the management of pesticide use.
Compact and very light at just 940g, the DV Wing is easy to use and can be launched by hand. It has enough battery power for autonomous flight times of 85 minutes and the on-board sensor is capable of capturing very high resolution images and highly accurate ortho-photos.