The Courier & Advertiser (Fife Edition)
Tech firms fuelling rise of autonomous field robots
Tom, Dick and Harry just a few of the mechanical marvels to watch out for
Research establishments and tech start-up companies around the world are developing field robots of all shapes, sizes and abilities.
For some, the principal objective is to bring even greater levels of targeted precision to field treatments in order to grow crops in a more environmentally friendly and cost-effective way.
For others, automation is the key goal, offering the potential to drive down labour and machinery costs by using equipment that needs no on-board operator and is light enough to eliminate soil compaction and subsequent remedial operations such as subsoiling. Offering a solution to the increasing difficulty of recruiting seasonal labour and skilled full-time farm staff is another objective.
Some will work individually, while others are being developed to operate in coordinated swarms, providing capacity from numbers rather than scale and others will collect data to help manage crop enterprises more effectively or perform tasks around the clock with consistent levels of performance.
In the main, autonomous vehicle technologies are best suited to highervalue crops – such as salads, vegetables and fruit – as well as others grown in a structured layout such as sugar beet and maize.
Small Robot Company
This UK-based business is developing complementary robotic vehicles that monitor crops and soil (Tom), provide precision weeding and feeding (Dick), and handle precision sowing and planting while recording data to allow treatment-per-plant precision (Harry).
Wilma is the “brains” of the system, using artificial intelligence software to extract information from crop models and data logged by Tom to create crop care prescriptions verified by the grower and implemented by Dick and Harry.
Co-founder Ben Scott-Robinson says the rewards of this technology will include precision crop treatments, reductions in pesticides, energy and CO2 emissions, lower production costs and improved soil structure from using light robots rather than heavy tractors.
Rather than purchasing the robots, the Small Robot Company proposes a leasing package for the specialised vehicles and software.
Carré
French cultivations equipment manufacturer Carré worked in partnership with Naïo Technologies on autonomous platform development and now produces its own four-wheel drive, four-wheel steer machine.
The Anatis uses satellite guidance and camera plant row recognition to guide adaptable hoeing elements mounted at the rear of the machine and the vehicle has a number of sensors to collect data such as ambient and soil temperature, humidity and soil moisture levels. It can also measure and report individual plant numbers and average size, all of which can help inform daily decision making on crop management and harvesting.
Deepfield Robotics
Public-funded prototype research in Germany has produced a commercial autonomous platform built by new Bosch business Deepfield Robotics.
BoniRob is a four-wheel drive, fourwheel steer vehicle with adjustable track width and optical or GPS guidance and interchangeable sensing equipment modules for in-field data recording and analysis. Crop research, automated field sensing and sampling, and remote monitoring are its initial roles, but Deepfield researchers are also trialling unassisted mechanical weed control that distinguishes weeds from crop plants and delivers a knock-out punch from a rod that rams them into the ground.
DOT
Developed in Canada by experienced manufacturers of no-till drills, the DOT platform operates field cultivation, sowing, spraying and spreading equipment for grain and oilseed crops.
Comprising a U-shaped four-wheel steer chassis with a Cummins diesel engine power pack and hydrostatic fourwheel drive, the device is smaller and lighter than typical farm tractor and implement combinations but gains capacity by being capable of operating around the clock subject to occasional refuelling stops.
Users create a path plan with obstacles identified for each field using a Windows Surface Pro tablet computer that sends instructions and receives operational data from the DOT unit. It can then be guided under manual control to pick up an implement and manoeuvred from the base to then operate autonomously.
ecoRobotix
This Swiss technology company has developed a spot-spraying robot that targets individual weeds surviving an initial overall or band spraying treatment.
Tipping the scales at just 130kg (286lb), the device has photovoltaic panels to charge batteries for the two electric drive motors and equipment for navigation along crop rows, weed identification and spraying.
A pair of multi-link arms suspended beneath the structure move rapidly backwards, forwards and sideways to momentarily position a spray nozzle above the target weed.
Each can span 2m (6ft 6in) or four rows spaced 30-50cm (12-20in) apart, enough to cover up to 3ha (7.5 acres) a day in sugar beet and oilseed rape, and in other crops once the detection technology has been developed. Fendt
Tractor and harvest machinery manufacturer Fendt is researching a small wheeled robotic vehicle that will operate in groups of maybe half a dozen at a time, managed through a computer cloud-based planning, monitoring and documentation platform.
The system will instruct each Xaver vehicle to follow the most effective route for overall efficiency and will compensate for one or more of them breaking down.
Designed initially for row crops such as maize, soya and the like, the robots will record where each seed is sown so that resulting plants can receive individual doses of fertiliser, crop protection treatments and maybe irrigation water.
These treatments could be fine-tuned across zones showing different levels of crop performance as detected by remote and in-field monitoring systems.
Naïo Technologies
French company Naïo Technologies is already in the commercial realm with its 1m (3ft 3in) long skid-steering Oz and larger-scale four-wheel circle-steering Dino weeding robots.
Oz is designed for hoeing in vegetable crops grown in relatively small outdoor or covered plots, while Dino is capable of autonomously seeding and hoeing four rows of vegetable or salad crops at a time using mid-mounted implements.
Both navigate by RTK GPS and cameras once the control system has been programmed with the fundamental layout of each field.
Wilma is the ‘brains’ of the system, using AI software to extract information from data logged by Tom