Robotic weeders are evolving quickly. Some zap their targets with lasers, hot oil or electricity, others target them with microscopic sprays or mechanical hoes. UK startup Earth Rover is betting that a carefully targeted pulse of concentrated light can do the job. Its lightweight solar- and battery-powered autonomous CLAWS robot—which can operate for up to 16 hours without a battery swap—combines weed control with crop monitoring, and is attracting interest from organic growers battling labor costs and conventional farmers grappling with herbicide resistance and a shrinking chemical toolbox. To date, the firm has sold its first 10 pilot units, with another 15 under construction, and plans to raise Series A funding next year to support further expansion. AgFunderNews (AFN) caught up with CEO James Miller (JM) at the World Agri-Tech Innovation Summit in London to find out how it works, what it costs, and whether the economics stack up. JM: The clue is in the title: Concentrated Light Autonomous Weeding and Scouting. So it is autonomous. It moves up and down the field, identifies weeds, protects the crop, zaps the weed with a small pulse of concentrated light, enough to boil the meristem and kill that weed off. At the same time, it can run with cameras on the front and back, so it can pick up per plant data for its scouting. So you can do crop monitoring, NDVI health Normalized Difference Vegetation Index, a widely used measure of crop health and vigor, crop position, and crop counting. AFN: Is pulsed light an established technology for killing weeds? JM: It’s a thermal method, and thermal methods have been used in this industry whether it’s burners using propane or whether it’s laser. Our subtle difference is we don’t use a collimated laser beam as such. We converge the light, much like a schoolboy would with a magnifying glass making mischief. And after that convergence point, it diverges. So within a set distance, it becomes eye-safe. AFN: How would it compare to what Carbon Robotics is doing with its LaserWeeder? JM: It has much less energy consumption, which enables us to be battery- and solar-powered. So we can run through a field for up to 16 hours with this particular unit here. We’ve run trials this summer without swapping batteries at all. AFN: What are the pros and cons of your approach versus other available options including hot oil, precision spraying, mechanical weeding, steam and so on? JM: I think there are a lot out there. The advantage of concentrated light is that it’s a lower-energy solution. So you’re aiming at minimum kilowatt hours per hectare. Because it’s autonomous, that’s your next big metric. You’ve removed the operator, you’ve removed the tractor, you’ve removed the diesel, and therefore your cost is really how does it get across a field fastest? And that comes down to kilowatt hours. AFN: How flexible and how robust is it? JM: It’s been developed for vegetable and salad crops and specialty crops. So bed systems, open field systems. Its track width can be adjusted from 1.5 to two meters, which is most of the bed sizes you find in that sector. Height variations for different bed heights, ridges on carrots, for example. So we can fit most of those farm operations. We have been asked for special ones, three meters wide or other configurations, and because the modules are exactly that—they’re modular—we can position those wherever we need to under a platform which is built around it. AFN: What’s the business case and the ROI for the grower? JM: The initial problem statement came from an organic farmer. So it was offsetting the cost of weeding. The only options in organic farming are mechanical, where you have to get in the field when it’s dry, or hand weeding. And there are big cost and labor availability challenges. So that’s where the original ROI came from. What we’re finding now is we’re getting more and more inquiries from conventional growers who are having chemicals removed from their use because they’re being banned, and they have no alternative for herbicide resistance in the weeds, to a point where the chemicals are no longer effective. So in goes concentrated light. It’s not affected by moisture and water levels in the ground. So you can go while it’s raining, while it’s windy, and at the same time you can zap a weed, and it doesn’t really care whether it’s herbicide-resistant or not. AFN: How precise is it, so that you’re only zapping the things you want to zap? JM: We’re down to millimeter precision, so we can get in very close to the base of the crop. The algorithms are now developed to a level where in onions, for example, we can see the emergence of that crop and not target that. We can target the weed right next to it. AFN: What is your business model? JM: We are selling the product. We’ve sold 10 pilot units and we’ve got another 15 being built over the next quarter, and they’re being bought outright by customers. We did think there would be a need for weeding as a service, but all farmers want to weed when they want to weed. So they want to own that equipment and run that through the field. The cost is around £150,000 with a discount for pioneers who come on board early. We’ve had farmers who bought them outright. We’ve had some who bought them on the back of projects sponsored by the supply chain and the food industry, and we’ve also had them funded as assets for farmers, so we’ve had a mixture. AFN: And what about the payback period? I guess that depends on the usage and the size of the farm? JM: It does. It depends on their weeding costs, and all farmers know their weeding costs! But this has had payback in anything from one to three years. AFN: How easy is it for you to scale up production of these robots, and how does manufacturing work? JM: The product itself is patent-protected in the UK and the US. The components are largely readily available, and we’ve designed this with farmers’ influence. Our chairman is a fifth-generation farmer, so it had to be repairable with baler twine, WD-40 a multipurpose lubricant, and a hammer! The steel structures around it are very easy to make and very easy to scale. The other components we’ve already been talking to our suppliers as to how we scale those in volume. The majority of that is off the shelf, so it’s already in mass production. The hardest part for us is to put it together fast enough. AFN: Is there a sweet spot in terms of the size of the farm where this makes sense? JM: I think that boils down again to the cost of weeding. It’s designed around a 100-hectare farm for growing a vegetable or salad crop. So that was the size it was designed for, roughly 60 weeds to 100 weeds a square meter as an average. We’ve had customers operating on as little as five hectares, 15 hectares, and they bought the unit because it’s cost effective. AFN: When were you founded and how much progress have you made since? JM: The company was established in 2017. We’ve been through a number of iterations of the technology with the scouting device, weeding solutions, developing the robotics, and only in the last four years, really, have we taken it through the alpha, beta phases, and what we see here: the pilot production model. AFN: How have you funded the business? JM: We’ve been very frugal, but it’s been with a mix of equity investment and some grant-funded projects through Innovate UK and the Forestry Commission through the Tree Production Innovation Fund. AFN: How are investors looking at this space? Is it more investable than it was two, three years ago? JM: Absolutely, so much so that we’re looking at our Series A for scale up next year, and looking at the next area of the product, beyond veg and salad into more broadacre cropping systems. Further reading: 🎥 Carbon Robotics: $100m+ revenue, new machine and a path to IPO 🎥 AgriPass on ag robotics 2.0: ‘We’re replicating the human weeding process but making it affordable at scale’ 🎥 ‘The whole system is now more investable’: Reservoir makes the case for ag robotics The post 🎥 Earth Rover bets on pulsed light to tackle weeds without chemicals appeared first on AgFunderNews.
CryptoAlerta — análise de criptomoedas e mercado em tempo real